Okay everyone.

We're gonna go ahead and get started.

We've got a form.

So I'll call this into order for three case meeting.

We've got a presentation from CPD and she should be able to leave your kicking this off.

Hi everybody.

So I think all of you know most of our command staff, but just in case you don't, we do have some new faces, and so um I'm gonna have them introduce themselves really quick and tell you what they're in charge of.

Yeah.

Uh Max doesn't have the sheet.

So I have the second time she's shown available, then I start by that one.

My name is Drake Correll.

I'm over with Federal Standards Bureau, and I have uh the Academy, uh policy, the PIOs, yeah.

Everybody wants Rick's job.

And I'm Mark Fischer, uh, so I'm over our patrol bureau.

I'm over our investigative bureau, so I have all our criminal investigation and also special investigations, which is narcotic street crimes, squat, etc.

Assistant chief hunter couldn't be here tonight, but he is over kind of our uh patrol support operations, so things like our park police, our homeless outreach team, our school resource officers, special events, things of that nature.

So we have this great online data tool that we've been working hard to refine.

I think there's still some refinements we'd like to do to it.

So feel free to jot down suggestions or things because we'll definitely take those into consideration.

But uh assistant chief is general is going to talk to you in depth and answer questions about how best to use the tool and a lot of the you're gonna hear hear the word nuance a lot.

So just be prepared for that.

And um, I think it your preference, but I think it's easy.

Yeah, it's they they want to choose a microphone.

It'd probably be easier if you guys ask questions as they come up because it is gonna get really confusing.

Probably if you don't councilman Foster, we have about an hour, a little under an hour for this.

Is that I wasn't given an amount of chance.

That should be the trick.

Okay.

So uh, because we can get as far deep into the weeds as you folks want to go uh on this.

Um, but please, I hope this will be a little bit more of a conversation than uh um really kind of instructional training deal.

So uh to start out with uh the the packet that I that are the papers I provided you here is is just summary of all of the talking points I kind of wanted to make.

Um I thought about just giving you kind of like cheat sheets to walk away from with just some of the definitions, um, but then I ended up just giving you everything so that way you guys have some notes to refer back to at some point if that's if that's valuable for you.

So um I thought we'd go through like kind of crime stats generally from the police's perspective, then get into uh the history of crime reporting because I think that that talks a little bit about why some of the way things are counted the way they are.

Um, go into our crime data dashboard, which is really the focus of this is uh I was asked to kind of give some instruction on how to use this crime data dashboard and to so you can look at it yourselves or give guidance in that in that regard.

Um, and then um just some kind of takeaways or recommendations I have with the crime data.

So the first thing I want to kind of just start out with by saying is that defining the terms is everything for us.

Um that when the terms are undefined, um, you're gonna get different answers to the same question because we're defining the terms differently.

So the easiest way or the best example that I think that I can come up with is our calls for service data.

So somebody says, hey, how many calls for service is the Columbia Police Department respond on in a year?

It's like, well, that's that's a great question, but calls for service is still too much of an undefined term.

That calls for service could mean calls where a citizen calls 911 or 311 requesting a police officer to come.

Um it could include self-initiated contacts, uh, which means like traffic stops, is the traffic stop count for calls for service.

Well, it depends on how you define the term.

Um, or some of the other things like our following up, evidence processing, um, report writing, some administrative tasks.

Does that count as for a calls for service?

And so when somebody asks the police department how many calls for service did you go on, um, and we say 90,000 calls for service in the last 12 months, 116,000 or 124,000, those all could be accurate numbers, but it defend depends on how you define the term call for service.

So another thing that's gonna come up with this data dashboard too is that there's conflicting priorities when presenting and examining data.

So the Columbia Police Department has hundreds of thousands of records um that we we utilize uh you know from a year year to year basis.

So the Columbia Police Department has hundreds of thousands of records that we we utilize uh you know from a year year to year basis, and if we just gave you all of those in an Excel document, like or a or a CSV file or some sort of database file, that's gonna be very hard to ingest and to examine.

And so when we publish the data in this data dashboard, we had to make choices on how to present the data.

And so the two conflicting problems with our transparency dashboard is the the issue of if do you map every offense or do you map every incident?

And we define an incident as like a single, a single event.

Um, so somebody, let's say we get called to a house and it's domestic disturbance, but there's also a child abuse um and drugs that are at that.

So that's one incident, but there might be three or four different offenses.

And so when we map the data, um, and do you map just the the incident or do you map the offense?

And if you map the offense, it might look like when you're looking at the map data that there were this a really bad street.

You know, there was 10 crimes that occurred here in the last day or the last two weeks or whatever the time frame that you're looking at.

When it when reality it was one police incident, and so what you'll see on the transparency dashboard is that we've prioritized incidents over offenses, and that affects the crime data as you're looking at it.

Um the other thing when it has to do with crime data when you're looking at it, is there's preferences and how to consume the data.

So when we we'll get deep into the or deeper into the weeds about what's a crime against society, what's a crime against property, and what's a crime against um a person.

But those are kind of the three broad categories for NIBERS or our crime reporter.

Yeah.

What do you mean when you said prioritize incidences?

What how do how is that prioritize?

So I'm saying prioritizing incidents over offenses.

Like we have an either or choice, right?

We could map all of the offenses out, you know, and and and look at the data that way, or we can map all of the incidents out.

We can't do both on a map.

You have to either choose are we mapping the offenses or are we mapping the incident?

So when you say prioritize, you mean each one of these flags is an incident?

Correct.

Each one of the flags on the transparency dashboard is an incident, and the highest call type.

Um, so if it was a homicide, uh rape and a drug offense, the homicide will be plotted, but the rape and the drug offense will not be on there if that all happened under one police call, one incident.

So does it is that yeah, am I because if we mapped all three of them, like an and it's kind of easy with three, but what if there was 10 um or 12 um and you know, offenses that occurred under a single incident, then it looks very differently.

So in this case, someone might be arrested for three offenses, but it's one incident.

Correct, correct.

Is there any way to cross-reference the offenses to be able to say to see where there is a prevalence of lower offenses, or is it purely incident-based?

Sure.

There, and there's certainly a way that the like the Columbia police department, the transparency dashboard is one shallow cut into crime stats.

And there's certainly different ways that us at the police department are interacting with that data to guide responses and to guide the deployment of resources.

And so this is the transparency dashboard that's on there is just one level of analysis.

There, that analysis tree goes deeper and deeper and deeper.

Um, and we'll talk about that a little bit more.

But there is there is, I think, improvements that can be made to this transparency.

We're the transparency dashboard that we have that's modeled through this software called Esri, um that is on the city's website.

This will this will this can grow more complicated um to it, but at some point, like I said, we're it already is so complicated that do you make it more complicated for like kind of end-level users?

Um, that's the question, really.

So the the different preferences and consuming data, and I think the example, there's two examples that I think come right to the top of the mind.

Um, one is robbery.

So robbery is classified as a crime against a person, or I'm sorry, it's not classified as crime, but it's classified as crime against property.

Um robbery at its very fundamental level is assault and theft equals robbery.

So somebody has a you know, through threat of force is acquiring property is what a robbery is.

And so there's a good question about whether or not you could classify as a crime against a person or a crime against property.

How NIBERS, the national uh incident-based reporting system has decided to classify that is uh under crimes against property.

And so with some of this, it's like we either need to the whole, and we'll get into NIBERS, but the whole idea of crime reporting is to be able to evaluate uh in an apples to apples way our crime data versus crime data in other communities, and that becomes impossible if we start kind of customizing how we're going to report crime or how we're ever going to share crime in in this way with the public.

And so there's a very compelling case to make robbery a crime against a person, but it's just not.

And so the preferences and how to consume data.

Another thing that we talk about is like homicide numbers.

So NIBERS, if there's if there's a double homicide where two people are killed in the same incident that's reported in a NIBER's as one homicide event.

Now we very much at the police department count homicides and the number of victims that there are.

But when we report homicides to the state and then to the federal government, uh, the FBI is the collector of crime stats that get to report it as a homicide event that has occurred.

Um so you know, the this is just kind of true of any kind of crime data, but but manipulating statistics is easy if the context is removed.

Um and that we can we can manipulate this data and massage this data to support any kind of political narrative that we'd like to, um, probably um by just tailoring your sample size and tailoring your time frames in which you're comparing.

Um, but just want to say that in general um before we move on.

And then I also want to say that crime data represents a moment in time.

Um, and that these crime statistics are a snapshot of when they were created.

And you know, my my counsel to all of you is that um you know that that talking points might not remain true for very long, especially the closer in time we look at the sample size.

If we're comparing two months data to the previous two months data, so we have our whole entirety of the look back is four months, you know, that stuff is going to be stale in a matter of days as opposed to weeks.

Um, and so my my recommendation to you is is that the longer your look back period, um, the more kind of robust that that observation is going to be.

Um that's it for kind of my my spiel on ground data.

We had some discussion.

We had some discussion, actually.

Um Jill gave us a report.

I remember this being before the most recent shooting downtown.

Um, but things happen so quickly that I'm not sure the exact timing.

But I think that there was some discussion about um an increase in staffing that um on downtown that related to an increase in self-reported crimes.

Um crimes against society and how it proactivity, yeah.

And I'm wondering if there's a way to um because you gave us a report with the numbers removed from self-reported entries from you know both data sets.

So we're comparing Apples Taples throughout our own database essentially.

It are we viewing that with an apple's taplos comparison in terms of that productivity self-entered in this database, and is there a way to look at you know, filter out those so we can look at that data with and without that thing, or or maybe reporting.

Yeah, I I think I understand what you're you're asking.

Is so we call we call that self-initiated activity.

So uh if uh an officer pulls over a car and um finds drugs in the car, um, as we are doing more traffic stops year over year, and as um traffic enforcement is increased, we have more, there are more opportunities for the officer to discover the crime occurring than there were if we had a lower staffing number, not as much time, quite as much time for productivity and things like that.

Um, it's gonna be very hard to look through at the transparency dashboard and make and make any kind of balls and strikes arguments um on that.

Um I think one of my frustrations here is exactly what you said that without context, it's very easy to misinterpret data.

Um, and on this dashboard, it's not clear what the context is.

And I wouldn't have known that that is an aspect of our data if it hadn't been presented to us, and also like you said, a week later, that report is now old and conclusions drawn may no longer be valid, but I don't know.

I can't view today's data compared to last year's data without that increased productivity kind of baked into it.

Um years from now that may fall out in relevance because we'll have a history of data to compare to.

Um it's very similar to if we undertake an initiative to crimes.

Yeah, which you might do.

Crimes that we know are underreported, like rape, sexual assault, right?

In the past, we've undertaken initiatives to try to increase reporting.

Well, was it the fact that more people reported, or was it the fact that there's actually an increase in sexual assault?

So it's it's the same thing you're talking about.

It's really hard to ever figure that out, but I think that's a big example.

The confounding variables looking at bottom line crime stats, the confounding variables are gonna be very difficult to suss out.

They're difficult for us to suss out as the command stuff for the police department.

Um, and we're in this data every day looking at everything that we can look at.

Um, and it's gonna be very difficult to get the context of that, um, which is why you have an excellent police chief that that can help provide the context um for the numbers.

So can I ask this question now?

Because it's no one will lose it.

Uh with all of that in mind, and how difficult it is to perhaps even draw conclusions.

What how do you all use the data that you are collecting?

And I know that you are to determine uh where resources are deployed, um where and and where you need to be putting an emphasis at any given time.

Yeah, I'll skip ahead, but um, no, I'm gonna do I want you to do that.

Um so there's two different things.

One from on on the patrol perspective is that uh um staff, I mean, staffing continues to be a challenge to Columbia Police Bond, right?

We are I I can't tell you how much infinitely better we are compared to uh let's call it pre-July um and post July, but it's still it's still challenging.

And so the the biggest crime uh or the biggest statistics that we interact with from a patrol level um are using it as management tools to uh evaluate hold officer productivity accountable to um balance workload, like the the low-hanging fruit from a statistics from my chair is in the management of employees um as much as it is for the for the for the guidance of deployment.

Um, if you look at this kind of crime stats that we have here, murder is just way too generic of a category to guide resource deployment.

Um, one because there's a bunch of different um homicides, or there's a bunch of different uh motivations for homicide, you know, over the course of the last couple of years, what we've seen is gang-related homicides trending down, or everything's trending downward.

Um, but this past couple years we've had three or four road rage homicides over the last two years, early shootings, um, and some of them have been ruled justifiable.

We've had a rise of domestic violence-related homicides as a uh um as a motivation behind it.

So while homicides are have really trended way down, the motivations for the homicides are different.

Um, and so by just looking at this like kind of bottom line murder number or homicide number, it doesn't, unless you're digging further into it to kind of look at the reasons behind this, it doesn't really help guide resource deployment.

What absolutely helps guides resource deployment is where are the shootings at or happening at.

Um, and we've started engaging with uh a report our analyst has worked out uh called Hot Streets, which looks at street segments um and where crimes are occurring.

Um, I wanted to kind of just bring up briefly touch on um Denver PD uses a model where they take a look at where past crime data has occurred or incidents have occurred.

Um they try to a certain degree, I don't know, predict where there's most likely to be kind of uh uh a rash of a specific crime, and they send officers there to do nothing but they call they their computer system generates a call for service, the officer gets dispatched to that street area, and then they are out on foot patrol on that street area during that time frame in which there's there's kind of past history is most likely to be indicative indicate that something's gonna occur.

The problem with the Columbia Police Mars, we just don't have the staffing capacity yet to do anything like that.

If the chief were to ask me to do that, I'd tell her how bad of an idea this is because we we're going from call, you know, a lot of times we're going or a lot of shifts we're going from call to call to call to call.

Um, and there needs to be more capacity to do some more data-driven deployments strategies like that.

Um would be would be my kind of response to that.

Um, so we are we are absolutely using the data to to guide resource deployment, um, but it's more nuanced than anything that'll be on this uh transparency portal.

And we're using a lot of other metrics like um that we're looking at officer productivity, like how many reports does an officer take per shift?

Is that close to their peers um in a on a per shift basis?

Um, you know, all the kind of things that you might do as a as managing employees kind of standpoint.

Um, all right.

Well, I want to talk a little bit about I've got a couple questions.

Um, and I apologize because these are super in the weeds.

Um, the first question is on the uh controlled vocabulary of the dashboard.

It sounds like you're using the NIBRS terms um for defining the crimes.

Um, but it looks like at least on crime trends, some of these there are a few different ways they get entered.

So like aggravated assault versus assault aggravated.

Yeah, let me or go ahead.

I'm sorry, I don't mean that interrupt you.

I the question is just like, is that intentional or is that we're we're starting in this and it will get figured out down the line.

So what you'll um we've we've had various compliance with the NIBER standard since 2018, October of 2018.

Um the the short answer is that in our profession, pick lists grow, they never shrink.

Because once a crime has been categorized as an aggravated assault, or let's say let's say let's use robbery, for example, I think it's better on here.

Once someone's classified as a robbery, then let's say the next year um from direction from a variety of of a variety of reasons, this could get directed to the Columbia police department.

About we need to start, we need to split up robbery, the NIVERS code of robbery, and we need to know if it's a if it's a robbery of a commercial convenience store.

We need to know if it's a robbery with a handgun, we know if it's a robbery with a strong-arm victim, and we start categorizing these differently now for a period of time.

Well, what do you how do you compare strong-arm robberies year over year?

When we just started looking at strong-arm robberies this year, does somebody go back in time and read tens of thousands of police reports to decide if this is a robbery or if this is a strong arm robbery or this is an armed robbery?

And they and the answer is like sometimes yes.

Well, that's what we did in the exercise in 2023, which is we had we had staff that reread all of 2023's police reports in order to correctly characterize crime data, which was a massive undertaking by that staff.

And so the the thing is is once something gets picked on a pick list, it's hard then in next year when we have this somebody has an idea of of hey, we need to we need to start measuring this.

And it's like, okay, but now do we how do we measure it against what happened last year, and how do we go back through all that data to re-categorize that data?

So at some point, there's naming convention changes or there's things, and this is the way that the that pick list shows the way that we have categorized data and brief here.

And when we get into this, I'll show you like some examples of of where you can kind of suss out when that changed or when that didn't change.

My next question was gonna be about neighborhoods, but I think that's coming up on your list of things to talk about, so I'll probably hold on to the code.

Yeah, neighborhoods is uh the neighborhoods in here, and and some of this was a canned report from this vendor, um software vendor, and so it uses neighborhoods, which we have we don't track um, we check addresses by by uh police department beat.

Um, and so the neighborhoods column on here shows the police department beats.

Um we can go through that here in a minute too.

Can we ask go?

Yes, yeah.

So when you're talking about the um classification for the pick list as the pick list becomes um more and more specific and more and more refined, um when you enter that data, do you have um or do you use um the ability to have um kind of a um generalized pick a top level category as well as whatever the specific one is, or as the pick lifts becomes more refined, do we lose the ability to track the larger group bait that it specified out of?

Yeah, that's a great question.

Um, and so if you go to this um the national incident-based reporting systems handout on the that I attached to this document, um, what you see is it's it's a loop, it's a list of offenses along with the offense code and whether that's a crime against person property or society.

And so if you look, um, let me try to find robbery here quickly.

Uh so like robberies on page, oh actually they're not page, the pages aren't numbered, but it's in alphabetical order.

You can see the robbery class goes to 12 there.

And so there's a top level code.

That 120 is the top level code, and then robbery strong arm, robbery with a handgun, robbery, whatever, all fill into 120.

And so it's really more of a problem when you start with 120 and then you try to separate things out than it is when you separate things out, they're still all feeding to 120, which is our top line kind of our bottom line robbery number.

Great, thanks.

Um, so NIVR's talked a little bit about it.

Um, but this really started out in um the UCR uniform crime reporting started in the 1920s.

Um, a group of police chiefs got together, and there was this this issue of how do we compare crime numbers?

Different cities are reporting, hey, our city's the safest, and then another city's going, no, our city's the safest.

Um the there had to be like the I wrote it's new on the top of this document.

I wrote it's nuanced, context matters, and how you count the beans changes the numbers, right?

How you count the beans changes the numbers.

And so they agreed on how the beans were going to be counted.

So this UCR was like the precursor to NIVERS.

Um, and UCR was a summary reporting system.

So where I was talking about before about how if you only count the highest level offense, that's what UCR did it only counted the highest level offense.

There's only one offense per incident.

Then the FBI and the International Association of Chiefs of Police got together and said, hey, we would like to look at all of the incidents that are in all the offenses that are in an incident.

Um, and we'd like those to be reported.

And so NIVER's born out of that.

Um in the I think NIVER started in the late 1980s, um, but it wasn't mandatory or compulsory until 2021.

Um that's finally when the federal government tied tax dollars to the states enforcing compliance with this national reporting standard.

Um so that happened in 2021.

We went live with it in 2018.

We were buying a new records management system at the time.

Um, and when we bought the new records management system, we made sure we were trying to future proof it a little bit, make sure it was NIVERS compliant.

And so we started NIR's reporting in October of 2018, um, but it didn't go live until October of 2021.

And what you'll kind of see is uh from 2018 through 2023, um, varying compliance with that directive and with that.

And so it was really a fundamental shift in our records unit, our records unit before current year, really, thought that you know, really so their mission was to memorialize the work product that the officers created.

The records kind of mandate and marching orders going forward is that they're responsible for ensuring the correctness of the data that they're managing.

So it's kind of two different philosophies of of how to think about what our kind of records bureau's job was.

Um, but it's a lot easier to get the six or seven people involved in that unit pulling in one direction than it is for the 145 officers that all kind of have different thoughts on things.

And so um in 2024 or the end of 2024, um, they read all uh the records unit read all of 2023's police reports, they made sure that they were they were coded correctly.

2024's reports are code of collectively 2025 reports are coded correctly.

And so um, I'd encourage when we're looking for an apples to apples comparison that we we really we look back to 2023, um, and before that, the data is I'm not saying it's not good, but it's not as good as it is post-2023.

Um okay, so now we're on to NIBERS.

We've talked a little bit about locally.

Um, yeah, the one thing I'd just like to really just kind of touch on too is that we really truly don't know what the underlying crime rate is of the city of Columbia.

We have no idea.

Um we can make some guesses at it, there's some metrics that we can look at towards towards it, but we don't know what's not reported to us.

Um I think that there's a that customer service absolutely pays plays a key into that, and our relationship between the police and the reporting population or the victimized population.

And so uh we can imagine, and we've seen it nationally that that immigrants are less likely to report being the victims of a crime than non-immigrants, and so um I think you know, we've we've talked kind of internally a lot about sexual assault in the homeless population.

What is that we know that crime that's occurring in some of these camps is way underreported.

Um and so we have there's an unknown, and then uh the customer service piece of it too is last year.

We our staffing was in such a state that that it was sometimes we had a larceny call from a like a shoplifting call from Walmart where the suspect wasn't in custody.

It might be three, four or five hours before we had a cop available to respond to take that report.

So with some of these businesses, the business is already done closed, and the employees have gone home before we have a cop that's available to take the report.

And so that report we get called in, it gets we get notified of it, but that never makes its way into the transparency dashboard because we don't have the full information that we need to complete a police report.

And so customer service, and then that group of there's a group of businesses that just stopped calling because the response times were so long for some of these kind of minor thefts that they stopped calling.

And so again, that is going to, oh you know, you might look, oh, we're down year over year for larcenies, but the context matters, right?

It's like we're down year over year, not because larcenies aren't happening, it's because the victims aren't reporting them because we don't have the staff to provide the service to get the data into the system to make it so we I know for sure that that was occurring, but to what extent it's it's kind of hard to tell.

Um, the other thing I want to say is that about that is that less than 10% of police calls generate a police report.

So, what the transparency dashboard is is a summary of information that officers have entered into our records management system.

Can you say that with more time?

Yeah, less than 10% of calls.

So we get a call.

Um, so like I told you, the calls for service numbers, it depends on how you define the term, it's either 90, 116, or 124,000 a year.

Um, but we take about 10,000 police reports a year.

So less than 10% of the calls for service.

So when somebody calls the police, less than 10% of the time, we actually complete a police report on that event.

Most of the time, it's handled without a police report.

Um, a lot of times that's because I mean there's a whole lot of reasons for that.

Um, people uh aren't reporting a crime.

Um they they're reporting something weird, but it's not necessarily criminal nature.

The officer investigates it and it's like nothing really has occurred here.

Um, or it could be child custody exchanges or somebody's got questions, or I mean, there's a whole lot of reasons why we have calls that don't result in police reports, but um the majority of them don't.

So we have this total unknown amount of crimes that occurred, a subset of that are ones that people call in about, and a subset of that are what we take a report on, and then only a subset of the reports that we take uh a subset of the reports that are entered into our records management system get reported to NIVERS.

We take a fair amount of info information reports where it's just like, hey, we're gonna we're gonna document this and memorialize this, but a crime hasn't occurred, and so those don't get reported to NIVERS because we're reporting criminal activity than NIVERS, not necessarily just all police reports.

Yeah.

Do we have a way?

Do we have a way to view the number of calls responded to compared to the number of reports filed mapped?

Uh I'm sure we can.

We don't um we don't look at you, you're kind of, I guess, getting into like where are we being inefficient to where officers are responding to, or what areas get the most calls versus what areas have the most incidents?

Sure.

Um and they may be the same.

Well, if you look at our beat map geographically, our beats are a combination of uh calls for service and criminal investigations completed, and we try to keep all of our beats roughly equal in the terms of workload but for officer, if that makes if that makes sense.

And so um, we just did a re-examination of the beats last year, and we found out they're they're pretty close still, and so we decided not to change them.

But um, 20 beat is got about the same number of total calls for service and total reports generated as 30 beat or 60 beat or or any of the eight geographic beats.

I was just gonna add something that we're working our way back to as staffing goes up as well that we use this data for that I think was helpful that we did years ago was um comp stat style management meetings, and so looking at we used to get a list of the top 10 locations for calls for service, and then we compared to that the top 10 locations where we took police reports, and then we would see where those things overlapped, and no surprise that all was on that top 10 list was all three Walmarts.

And I'm not picking on Walmart, it's just there it's well known, those are locations where we get a lot of calls, and there's police reports that get taken.

So then we look at those things overlap and go, okay, how do we influence this?

How do we drive that number down?

Can we get them to report shopliftings online?

We try to come up with creative solutions to be more efficient.

Um, or are there ordinances that we need to look at, just different things.

So we're working our way back towards that.

That's actually a huge piece of tomorrow is the kickoff of our new record management system project.

So that's one of the things we're most excited about, or I am, I guess, coming out of this is how we're gonna be able to better use data in the system.

So I just wanted to.

Yeah, no, the analytics piece of this new records management.

Our current records management system has virtually no analytics um to look at, virtually none.

And so um that was a big need for us is getting analytics um in the officers' hands to look at their geographic areas responsibility, which we're excited about.

This uh this new product.

Um one of the biggest check marks in it was that it had it has the I think the most robust analytics um out of any police records management system available.

So uh so we'll get into the meet and tails here.

Uh so uh the easiest way for me to get to the dashboard is to get it to it from the Columbia Police Barnes website.

So if you look at the those crime data dashboard, um that's where we're what we're really talking about.

And so what you see here is the is the crimes against person, crimes against property, and crimes against society um on this summary page.

So the data in the dashboard goes back to 2023 and it's live up to three days ago.

So we have a three-day window where where it's live.

Now, one of the things that I'll caution you on looking at is that the last 28 days data in this website as a whole is the least accurate for a bunch of different reasons.

Um, one is that it goes off of the date that the crime was reported to have occurred at.

And a lot of times we might not get notified that a crime most time we get notified that a crime occurred the same, the day it occurs, but a lot of times we don't get notified for a week later.

And so, by the very nature of trying to make this as live as possible, we're also saying that the last kind of month is inaccurate.

If you had a burglary that your house that you discovered a week uh later, and then you call the police and you got them on your surveillance camera, and we have a date and time of when the burglary actually occurred, this data will go back and update, you know, and add one to the burglary count from the week before.

But so that's why I'm saying the last 28 days is the least accurate.

Um what this this summary slide is missing.

Um, and this as this evolves, we're gonna we're gonna add this in.

Uh is it misses the it's missing what we call the group B offenses.

And if you look at the NIVERS um pamphlet that I attached to this later, you'll see that group A offenses are most of the just kind of the heavy hitters of what criminal activity is.

What it does not include group B offenses on this summary page, and I wish that it would.

If you go into both the public crime map and the crimes trends dashboard, we can look specifically at the group B offenses.

Um, but this summary page does not include the group B offenses.

So my ask of what this will, this page will hopefully look like in the future is to add a block down at the bottom here with the group B offenses.

And the group B offenses are and they're listed, I think on the last page there.

Um bad checks, um, vagrancy violations, disorderly conduct, driving over the influence, drunkenness, family offenses, liquor law violations, runaways, peeping toms, trespass, and all other offenses.

And so the group B offenses, let's go ahead and look at those right now.

Um, I'm gonna go into the crime trends dashboard.

And under this NIVERS code here, I'm gonna oh no.

Crimes against, I'm gonna go into this um group B offenses.

And so I think uh councilwoman Carroll's point earlier about uh if you can see this year-to-date column.

We have year to date, we've had 25 and 40 arrests for group B violations compared to a year ago, year to date.

So January 1 through 11 through, well, it's actually it'll actually be 1031 of 2024.

We have a 59% increase on group B offenses.

So when we're talking about the increased enforcement that we've been doing downtown, we also had another initiative where we've we've increased enforcement in the Connolly Road Corridor.

Um things like trespassing, liquor law violations, um, all of the the kind of proactive quality of life, misdemeanor, crime enforcement.

That's gonna be a huge driver of total crime rate this year.

It's gonna be the thousand more cases and arrests we made for these group B offenses.

So the group B offenses only get counted on the arrest, they don't get counted.

Um so if we don't, if we take a report on trespassing, but we don't make an arrest, we don't report that to the state or to the FBI.

We only make the report to the state of the FBI as if we've made an arrest.

Um, and that's just part of the NIVERS rules for fores that does the group offense stuff make sense.

Um I know that this has been kind of a can you run through what those are again exactly.

Yeah, so they're on have a better idea, and and I guess I'm wondering why we didn't have that in that summary page.

Yeah, you know, that's uh it's uh this the this is a new it's ongoing.

Um we should we should we're gonna continue to tweak this.

Um but yeah, the uh group B offenses, and some of these offense titles don't make a lot of sense for Missouri, but remember this is a national, you know, we don't have uh a loitering charge as been very publicly commented on uh in the last couple weeks.

Um, but there is a loitering charge on this group B offense because they have to make it a way that all municipalities and and agencies in the country can report to it.

So the group B offenses are bad checks, and in Boone County, bad checks are actually investigated by the prosecutor's office directly.

Um there's a report you make.

Um so if we get a bad checks call, we'll we refer them to the prosecuting attorney's office.

Um you have curfew loitering and vagrancy violations, disorderly conduct, driving under the influence, drunkenness, which we don't have a charge for, family offenses, liquor law violations, peeping toms, runaways, um trespassing on real property, that's gonna be a huge driver of this number.

Um, and then all their offenses, and you can see they're all coded to 90.

So all the 90 offenses are coded to this group B offenses.

And you said they only get reported if there's an arrest.

Correct.

Is that if there is an arrest for this offense, or if there's an arrest for something else and then it's also a group B offense, it would get reported.

No, an arrest for trespassing by itself would add one to this number that we're looking at right here.

Um, I know if I've kind of jumped around here too.

Um, I will say that on the data dashboard.

Um, we'll go back, I'll go back to the summary page.

Okay.

Um if you when I say that the the last 28 days is least accurate, every one of these graphs is gonna end in a downturn far to the right because the the crimes are never when I say they're inaccurate, they're never gonna be overreported for the last 28 days, but they're almost always going to be under-reported for the last 28 days.

So all of these graphs are going to be showing um down this one's sideways, but I mean that number right here, 33 for the last 20 days is under-reported, not over-reported.

So am I understanding that what you're saying by that last data point being underreported, is that as you proceed, it may very well go up once you enter in crimes where they determine the date on afterwards.

Yeah, and it's that's one of the reasons it's underreported.

Another big reason is too is that um crime or the reports have to be reviewed and locked by several layers of review prior to that data being included on this dashboard.

Um, and so the officer this has to write the report, the supervisor has to approve the report, and records has to approve the report.

As part of making sure that this data is accurate and records change in mission from being focused on just memorializing the officers' work product to being responsible for the correctness of the data that was been included.

There's a records review, and so between the office, the the officer supervisor and the officer of the officer supervisor and the records review, sometimes it can take longer than three days.

Three days should be the case, but if we have a big event that occurs, um, like we have a homicide late at night or some sort of significant event late at night, like none of the reports from that shift are gonna get done that shift because the officers already here multiple hours for overtime just to work that event.

Okay, uh I know if I'll spend a lot of time talking here, but uh we'll go real quickly through the uh um the public crime map.

So we talked about the list to your point.

We already talked about that the lists in this profession grow and they never shrink.

Um, I was actually looking at crime call types from the early 70s, and one of the crime call types was annoying females was the name of the crime call.

And it wasn't that the females that were annoying, it was like men and cat calling at women on and so the the but anyways, that was a crime type in the early 70s, and so we don't have a crime type of annoying females anymore, but but that crime type would be classified under something else now.

And so now if you were to look at with whole completely different crime types from the 70s versus now, how do you make an apples to apples comparison?

And it's really difficult.

I have a crime type question.

Yeah, um, I'm seeing on here um under the heading crime types, overdose um is overdose its own crime category, or is that being reported for another reason?

I I don't it overdose locations are something that we've tracked.

Um I don't know if that's a NIBER's reportable defense lance.

Yeah, no, there's no controlled substance is an exception uh you know, but that makes it not even arrestable if it's in the state because everything is reporting existing, but they're using a controlled substance in the commission of the overdose.

There was a crime that occurred, right?

It's not a restable if they are seeking, it might be it's on the map, yeah.

It might be reported under, I'll have to get back to you on that because it might be reported under possession of a controlled substance.

Yeah, like it might be a 35A violation, and I know there's a lot of there's a lot of data.

Sometimes it's like you said we we count the beans differently every time.

Yep.

Overdoses, obviously, everybody knows this was a big issue, you know, over the last five years because we saw such an uptick in your early 2020s, and so there's a lot of data and collection points that were to be gap captured.

So they could have added a point on here.

Yeah, and that was kind of what surprised me because there's only one listed.

So I was thinking if we were tracking this, there would probably be so we actually land on a scooter code.

I should have done that sooner.

This is not one of the desktop 28 days.

Oh, okay.

So uh the the crime map defaults to the last 28 days of crime, so that's why you only see the 571.

And again, that 571 is likely way underreported than the actuality of it.

So the the one thing I'll say, uh is it on?

Okay.

Uh the one thing I'll say about overdoses.

Overdoses are uh a very difficult thing for us to track.

Um there could be an overdose investigation that leads to a arrest for murder second that will be categorized in the murder offenses when we reported to Nyvers.

There could be an overdose that results in a distribution of controlled substance.

So it could hit several different categories.

So uh I don't know exactly the the background on that.

I do know it's a it's a data point we do track internally at the Columbia Police Department.

We actually have a separate database we created just for overdose tracking.

So I don't know if I clarified that or confused the question, but there's a lot of ways you can count count an overdose investigation.

So it did clarify, and it sounds like it's just a intro onto a new database problem.

Uh, but I would love to see it not listed under the heading crime types.

I just want to draw attention to the time.

Um we can go a little bit past six o'clock update, but but uh just want to keep that in order in terms of yeah, I'll I'll wrap it up in five minutes or so, and then there may be plenty of time for questions.

The one thing I I'd like to tell you about the the police crime map when you're working on this is if you zoom in to an area of the map, the crime total over here, it only shows you what is in the screen mapped.

So that can uh it certainly has confused.

I feel a couple questions about that.

Uh so these filters up here will let you will let you kind of dig into this um a little bit further.

Um you can you can filter the dates that you're looking for.

So if you wanted to create two different reports that compare year over year, um the NIVERS offense category, um, that's right here.

Um you can look at weapon type and case status.

Uh yes, ma'am.

You talk about the laboratory six that are easy wall excuse.

Yeah, it's so whenever we're looking at crimes in neighborhoods.

Um, we often pull out crimes that are at 600 East Walnut.

So if you look for again how you count the beans.

So when joint communications receives a 911 call for service and they don't know what location it occurred in the city, but they think it occurred in the city, they throw it at the Columbia Police Department that this crime occurred at the Columbia Police Department.

Um, and so because we have to have an address because everything's tracked by addresses here.

And so when you're looking at crime data downtown specifically, you got to be really careful about whether or not you're looking at calls for seven 600 East Walnut or you're not looking at calls for 600 East Walnut.

Um, because we always pull that back out to to look at at data because we don't know that occurred there.

Um we have a way to pull that out when we're viewing the dashboard, probably not.

Probably not is the is the answer.

All the addresses in this dashboard are generic to the block range, the street segment.

So we're we're not putting the specific address out there.

Uh so the crime data dashboard, um, very quickly, or the crime trends dashboard, I'm sorry.

We talked a little bit about the great B offenses that we can that we we picked on a slider over here.

Uh the other thing I'd like to tell you too is, and there's not what the other request that I'm having to update our the summary page is to include a violent crime summary.

So violent crime is something that I think that everyone in the community is kind of concerned about.

Um, but what does that mean when you actually look at what is violent crimes mean?

It's not crimes against persons, but uh the NIVERS definition of violent crime includes murder, non-negligent manslaughter, forcible rape, robbery, and aggravated assault.

And so if you ever wanted to look at what's our kind of crime data um or where we're at violent crime wise, uh, you could go into these NIBERS codes and pick those specific entries to look at our our violent crime rate.

So I wanted to give that definition there, which is a bullet point under this uh crime data dashboard.

Um, the other thing that I'd like to just say with this too is um if you look at our shots fired numbers and you look at just the category of shots fired um on the crime type, that's not gonna show you every incident where a gun was fired in Columbia because if it if it was a homicide, um if it was a rape, if it was an aggravated assault, like some of those shots fired are gonna be in these other crime categories, and they're not gonna be double counted both in the shots fired and the other crime category.

And so um, I'd like to just kind of guess I guess go towards my recommendations here, which is when you when you interact with this data, my encouragement for you folks is to is to interact make your searches as broad as possible.

So if you look at the and we could look into robbery here specifically, but you can it's very clear here when there was a change in the robbery coding definitions, because you can just toggle like you can toggle the box on and off and be like, oh, in 2023 we had no robberies.

Well, that's not true, right?

It was just toggle bundles or other things.

So when you search for robberies, search for all of the robberies, and the larger the date range you can do, the better.

And so with these shots fired, when the the chief was making some statements about how shots fired were down year over year in response to some of the inquiries that she was having.

Um I think she gave the numbers of in 20 year to date in 2023, we had about a hundred um shots fired cases, and year to date in 2025, we had about 50.

And if you went and you look just at the data on the dashboard, you would have seen something like 96 and 48.

Like the numbers weren't exactly the same, but the trends are gonna be apparent, um, even though the the specific counts might be one or two off, depending exactly on how you're counting the things.

And so I'd be I'm encouraging you to uh um make your searches as broad as possible, both in time and crime categories.

Um and be less focused on matching exact numbers and most more focused on on broader trends.

Uh and then the other thing too is that a big thing that's gonna make these numbers go up or down is what is police staffing?

Um and what how many resources do we have available to to throw?

What are our clearance rates?

You know, are we if we're making more arrests, or are we making more arrests because we're solving we have a higher percentage of crimes being reported to us, or is it because we're solving higher percentage of those crimes?

Like those things are gonna be like unclear um just on a base look at the at some of the data, but the confounding variables are are very real, and I'd be I'd caution you against making causation statements um based on the data.

Okay, all right.

Thank you very much.

Uh Magic, there's two questions.

Yep, yep.

Um this might be also for two Chief Slug.

So in the conversations that have been taking place most recently within references to the data, there's been uh disagreement about what the data means.

So what conclusions are you all coming to so far?

And I'm thinking about this, particularly in terms of what actions the council might make.

And among the actions that we've taken over the past few years is to address the need for more police officers by increasing pay, establishing our own uh police academy and so forth.

Uh but moving forward.

How might these numbers guide what council does in terms of policy?

So I think that you know these group B offenses are absolutely driving up the total crime rate, you know, as reported.

But that's not that's not the underlying what's the the question I'm most concerned with answering is is total crime down or up.

Um and that's the unknown number is that down or up.

Just because we're not making the arrest for trespassing, doesn't mean the trespasser didn't occur.

And so I'll say that you know what these numbers show is that you know, crimes against the violent crime is up year over year.

It's up, I don't know, it's something like 15%, but that's being largely driven by aggravated assaults and largely driven by domestic aggravated assaults, and so our homicides are were were very low compared to you know year over year 11 to 3.

Um, and so the traffic fatalities are very low year over year, but some of the some of this data is I mean, crime like total crime reported crime is up.

Um but what what's the the conclusion from a policymaking standpoint?

Lance can have your answer.

Yeah, yeah, absolutely.

The only thing I I'm not sure that's working, yeah.

I don't know anybody else in the movie, but yeah, that's fun right there.

That would be this.

I'm just gonna jump in here on the on the data points where I think it really can affect you guys.

Um, if you look at our benchmark city stuff, one thing we do in the benchmark cities we report all these crime data to our benchmark cities.

We also report all the demographics of our police agency to the benchmark cities, and what we're able to do with that, we're kind of able to compare where we are at to our counterparts, right?

Because when we find cities that have equal rates of violent crime, property crimes, crimes against society, equal size and demographics, and their police agency is 270 people, we're like, okay, we need to expand the size of our police agency, right?

If their budget is 20 to 30 percent larger than ours, we say we need to expand that, or if we're on point with them, we say we're appropriately sized, we try to benchmark ourselves against comparable cities, and crime data is a big big component of that because if we take 120,000 people and we compare it to I don't know, pick the highest crime area in the country.

That's not an apples to apples comparison because they're having 90 100 homicides in that same area, you know.

So I think the crime data where it's important for us is when we look at those benchmark cities and those indexes, they able they're able to identify comparable agencies that we should really be looking to to see if we're consistent with our counterparts that makes sense, yeah.

So yeah, I mean my biggest thing is don't let the the context of the the minor offenses increasing because our customer servers increasing because our relationship with the victimized population is is allowing for more of these people to be reportable.

Don't let don't let that um influence like we're we're going in the right direction with enforcement, we're going in the right direction in these areas.

Um, but I think the data can be manipulated into making it look like we're not.

I think one of the most important points Mark made on that topic was we as a command staff, we have great concern that things have become more and more underreported the longer we've been understaffed, if that makes sense.

Now, how can we ever prove what's been not reported?

Well, we can't.

Um, but we do know there's been a level of frustration over the years with the customer service we've been able to provide.

So we have felt like we call these I feel statements when we have meetings, but if people don't feel like they can get a response and a timely response to something, they just stop calling.

I mean, that that's a pretty common human response.

And so as we've not been able to respond to things as quickly or even follow-up, right?

I mean, if you can't ideally an officer that's dealing with any type of crime or calls for service in there beat, should be able to follow a call through till it's logical end, whether that's on the same day or they come back their next shift and do some follow-up and some you know investigation, things like that.

That's just been nearly non-existent when they were starting a shift with 10 or 12 officers.

So as that capacity grows, we're hoping people will get back in the habit of feeling comfortable and like they're going to get a response from us from reporting things, and we're going to work that case through to its logical conclusion.

Lance, of course, loves when we don't refer everything to detectives, but there was a time when it felt like that was the only option, but it wasn't a great option because he's just as understaffed as patrol is.

So you know, it's it's that you know, apples and oranges type conversation that we get into, but when you have 10 or 12 officers on a shift and they don't have time to go pick up videos the next day to do follow-up or go get a piece of evidence or go interview a witness, all that gets pushed off on detectives, and I told you guys in other meetings.

There were times where we had one person assigned to all the property crimes in the city.

So what level of service are we providing?

So people just kind of give up sometimes, I think.

So as we build back that capacity, I think we're gonna see another increase in reporting just because we're able to be more responsive.

Um, how do you track all that stuff?

I'm not sure that you can.

Um, but I'm not going to be surprised if some of these numbers go up the more we get staffed.

Okay.

Anything else?

All right, thank you so much.

Gives us a lot to think about, and I'm uh really gonna encourage everyone to do what I intend to do is go back and look at your summary here and refresh our minds with that.

Thank you.

I was gonna ask if we can have a digital version of this.

And also actually, um, if this kind of summary information could be made available to the public, I think that would be really helpful for this.

Um, one of my concerns is that it is nuanced data, and um, it can be interpreted any manner of ways.

Um the context is not apparent on the dashboard.

Um, so it can be misused any number of ways, whether or not the person is intending to do so or not.

Um, so you know, maybe having guides and disclaimers available to the public will help people be aware of where the pitfalls are.

Yeah, the uh uh and I put it in the summary document here, but the there's gonna be some changes um to the to this dashboard.

We we started out with which was more or less the canned report from Ezri, which is the vendor on this.

Um, but there's a ways to go to make it more transparent, more user-friendly, and explain a little bit more of the things.

So um, I think to the point of like some of the disclaimers, some of the have some of this on here is absolutely the intent over the next you know, over the 20th.

Thanks.

That's a good point because we really want the public to be able to use this and to draw the right conclusions from it.

All right, thanks.

I'm gonna uh offer a motion in a moment to go into closed session and just say that we'll take a quick break in case anybody wants to go do something in between that.

So I move that the city council of the city of Columbia, Missouri immediately go into a closed meeting here in conference room 1A1B of City Hall to discuss confidential or privileged communications between a public governmental body or its representatives and its attorneys pursuant to section 610.0211 of the revised statutes of Missouri and operational guidelines, policies, and specific response plans developed, adopted, or maintained by any public agency responsible for law enforcement, public safety, first response or public health to use in responding to or preventing any critical incident which has potential to endanger individual or public safety or a health pursuant to section 610.02119D of the revised statutes of Missouri is required by section 610.202119D, the city council hereby declares the disclosure of information to be discussed in closed session pursuant to this section.

The public governmental body's ability to protect the security or safety of persons or real property and that the public interest in non-disclosure outweighs the public interest in disclosure of the records gonna have a second.

Sorry.

Thank you.

Uh and I'll call roll now, Zelwood.

Yes.

The sample.

Yes.

Foster, yes, Ms.

Waterman, yes, Ms.

Peters, yes, Ms.

Buffalo, Miss Carroll.

Yes.

All right.