Monday, October 31, 2016

Using Google Alerts for Crime Analysis

I originally wrote this post for a software company’s blog in 2014. This company was bought out and recently their blog and website have been removed from the web permanently. I am reposting it here for posterity. 

I might be showing my age but I can remember a time before smartphones, before the Internet, and before home computers. I can remember searching card catalogs at libraries that were actually drawers full of cards.

I can also remember reference publications that indexed magazine articles in big dusty volumes.  If you were searching for an article about a particular topic you could search through these books and find a list of magazine articles that contained them. It always seemed to be my luck that the library I was at never had any of these magazines on their shelves. However, they could get me a copy though an inter-library loan in a few weeks. So much for timely research.

Search engines like Google or Bing have turned this on it's head. Now, anyone with a web browser and an Internet connection can query a search engine and find relevant articles with just a click.

Previously, I covered Improving Your Searches With Boolean Operators In that post I looked at ways you could improve the likelihood that your search will find the information your looking for.

As powerful as Google is, wouldn't it be nice if it could keep a watch on the Internet and let you know anytime it finds the search terms you specify? After all, isn't automation what computers are good for? Actually, it can with a feature called Google Alerts.

Google Alerts lets you specify search terms that it will monitor Internet sites for and when it finds them it will either send you an email or it will put them into an RSS feed that you can monitor using an RSS feed reading application.

Let's look at an example of how I use Google Alerts in my workflow. I like to monitor the Internet for news stories that mention the community where I work, Killeen, Texas. It's nice to have a heads up when these stories come out. While I already monitor most of the local news outlets via RSS and Twitter, there are also mentions of the community in other news outlets I wouldn't normally monitor.

For instance, I got a Google Alert email that indicated my search terms of "Killeen, TX" was used in a small town news website in Indiana. When I followed the link to the story in the site the quoted news article was about a recent highway drug interdiction arrest where a resident of our fair city got nabbed with a load of drugs on his way to points north.

The agency that made the arrest had not thought to contact us so I looked them up and gave them a call about this person. I was able to give them information regarding cases we'd previously worked on this person so they could make this information a part of their investigation. I was also able to add the information from their case to the intelligence information we held on this subject.

Here are some ideas for Alerts to set to get you started:

  • Your agency name - to catch news stories about your department
  • Crime M.O.'s such as "metal theft" or "bank robbery"
  • A crime analysis tool such as "GIS"
  • If your agency is working a high profile case, the name of the victim or suspect
  • And of course "crime analysis"

You'll need a Google account if you don't already have one. Once you establish that, head over to the Google Alerts page to get started. Once there you'll be given a set of options. You can enter your search terms, pick the type of sites you want to monitor such as news, blogs, discussions, etc. Then you pick language, region, periodicity, and how you want to get the alerts.

A neat feature is that once you enter your search terms, you'll see a sample of the types of entries you'll get with that search term. This works really well for modifying your terms in real time to see what type of results you will likely get. Once you enter an alert you can also use the Manage Your Alerts option to modify or delete alerts.

What types of alerts would you likely find useful?

Monday, October 24, 2016

Using Mind Mapping to Brainstorm Ideas

I originally wrote this post for a software company’s blog in 2014. This company was bought out and recently their blog and website have been removed from the web permanently. I am reposting it here for posterity. 

Becoming a crime analyst was quite a change from being a police officer. As a police officer my tasks were being directed largely by someone else. A Call For Service gets you dispatched somewhere to start an investigation, assist a citizen, etc. As a detective, your work is often directed by others. Your supervisor assigns you an investigation and you then follow your agency's investigative protocol.

My experience as a crime analyst is much different. While you still get tasked by others who may request information or a report on crime trends, other projects require more self directed work. One of these types of projects that requires more analytical thinking is developing solutions to local crime problems.

As an analyst I may begin to see a local crime trend develop and then it falls to me to analyze it and brainstorm a solution to that problem. To do this well often requires a way to formalize your thinking about the problem and it's possible solutions.

One thing that I found to be a great help for this kind of brainstorming was something called Mind Mapping. The term was coined by British author and psychologist Tony Buzan but in essence, mind mapping is a way to visually diagram information.

When we were in school, we were often taught to create an outline when we were researching a term paper. Depending on the requirements of your instructor, an outline tends to follow a hierarchical format. While I love outlines, the structure they have is linear. It starts from the top and works it's way down.

However, when you are brainstorming an idea your thoughts don't always come to you in a linear fashion. While a hierarchical outline works well when you start at the beginning and work towards the end of something, the format doesn't as easily lend itself to starting in the middle of a problem and working outwards, or starting at the end and working backwards.

Mind mapping is a diagram that works like a non-linear outline. You define the topic in the middle and then work outwards. Because of it's non-linear format I find that I am less constrained to start at the beginning. I can start wherever my thoughts take me and work in whatever direction I need.
For instance, let's say I discover a trend of increasing vehicle burglaries. I'd start off by putting the words "Vehicle Burglary" in the middle of the page. Then as I think about this, I know that there are proactive and reactive ways of dealing with a crime problem. I'll add branches to the main topic for Proactive and Reactive. Then I'll continue to branch off them until I cover all the bases.

Because a mind map is non-linear, I am not constrained to work in any particular direction. If an idea pops up, I can add it and move on capturing all my thoughts on this topic.

You can see a sample mind map I created for a solution to Vehicle Burglary here.



There are a number of ways to create mind maps. The first and likely one of the easiest is with a pen and paper. There are a number of advantages, its cheap for one. The other is that you can work with a lot less distractions.

Computer applications whether an application installed on your machine or one that is web based is the other way. You can generate some really complex diagrams that are neat and legible. Another advantage a computer generated mind map has is that it's easy to rearrange your diagram or to move whole sections around. You can also link to other computer based resources such as using Internet hyperlinks. The downside is cost (although there are some very good free and open source applications) and distractions.

Brainstorming while mind mapping should be more about getting all the information down quickly so you can understand it. Mind mapping like outlining is just a tool. If you find yourself spending more time playing with the tool than using the tool to create you might be missing the point. The website Vox recently had an interesting article titled Why you should take notes by hand — not on a laptop that gives credence to the idea that people learn and retain more when they take notes by hand rather than using a computer.

In addition to brainstorming, there are other things that people use mind maps for. The most common are note taking such as in a lecture or a meeting and task management.

The website WikiHow has a great tutorial on How To Make a Mind Map that will walk you though the basics.

So how do you brainstorm ideas? What do you find works best for you?

Monday, October 17, 2016

Modus Operandi Can Help You Identify Serial Offenders

I originally wrote this post for a software company’s blog in 2014. This company was bought out and recently their blog and website have been removed from the web permanently. I am reposting it here for posterity. 

We've often heard of the 80/20 Rule. In law enforcement, the rule says that 80% of your crime is committed by 20% of your offenders. While the exact percentages are probably not 80/20, it is safe to say that a significant amount of the crime in your jurisdiction is committed by a minority of offenders.

It would also be nice of these prolific offenders if they would identify the crimes they are responsible for. It would make linking them together a whole lot easier. We can't expect offenders to leave their business card on the door or to sign their work. But in one way, they do. Often times, their "Modus Operandi" can link them to the offenses they're responsible for.    

So just what is Modus Operandi? The phrase is Latin for "method of operation". Often times we hear it when referring to serial killers or some other exotic crime. But it doesn't just apply to exotic crimes, every crime and every criminal has a modus operandi or M.O. A criminal's M.O. can help you determine what crimes he or she is responsible for.

A prolific offender often develops an M.O. over time. Criminals will learn as they begin committing crimes. As they become more successful and more comfortable in committing offenses, they will usually stick with what works for them.

Economic principles also apply to offenders. If they have to spend more time and effort to commit a crime than the payoff, they will eventually abandon that method and seek out one where the economics are more stacked in their favor. They will learn from others in their social circles or even through trial and error. Eventually, they will gravitate to something that works.

If your agency faithfully records M.O.'s in your offense reports, you can use this data to identify offenses that can at times be attributed to specific offenders or groups of offenders. An offender identified in one case, can be linked through M.O. to another case where he or she was not known.

I worked on a serial armed robber whose M.O. allowed us to tie him to quite a number of armed robberies. He wore similar clothing during each robbery, he targeted similar businesses and made specific comments to the victims in each crime. At the beginning of the series started targeting these businesses in the late evening hours. However, he shifted to hitting businesses during the early morning hours and was more successful. By reviewing armed robberies in our jurisdiction and in neighboring jurisdictions we were able to pick out crimes he was responsible for because his M.O. had been unique and had been recorded.

It's important to try and identify as many crimes in a crime series as possible because the more information you have, the more likely you are to develop enough information to be able to solve a crime series. In fact, if you identify enough crimes in a series you can then start to predict future incidents based on statistical techniques such Mean Days Between Hits or other more sophisticated analysis.

Here are some general M.O. areas you need to try and cover in your analysis.

  • Journey to crime scene
  • Method of entry and entry point
  • Actions inside
  • Weapons used 
  • Words to victim
  • Actions towards victim
  • Items taken
  • Method of exit and exit point
  • Flight from crime scene

It's important to encourage your officers to gather M.O. information carefully during the initial investigation. You can further this end by taking opportunities to develop training materials that provide information on gathering M.O. information and sharing success stories where analysis of M.O. was important.

It's also important to identify known offenders associated with particular M.O.'s. If their M.O. is unique enough, and you begin to have a number of crimes with that or a very similar M.O. this  will give you a place to start on the suspects list. For instance if you know a particular offender is associated with rooftop business burglaries, then you can start looking at him a bit harder when you start having a rash of rooftop business burglaries.

Another often times fruitful endeavor is to debrief offenders about their M.O. or their knowledge of other criminal's M.O.'s. If your detectives can develop a rapport with an offender you can learn a lot from a good thorough debriefing. Questioning them about their M.O. can give you insights into their target selection allowing you to develop information that can guide further investigations or crime prevention efforts.

How does your agency collect and analyze modus operandi information?

Monday, October 10, 2016

Improving Your Searches With Boolean Operators

I originally wrote this post for a software company’s blog in 2014. This company was bought out and recently their blog and website have been removed from the web permanently. I am reposting it here for posterity. 

Much of the foundations of computer science where laid in a different era, one long before smartphones, laptops and desktop computers were even a dream. A major part of this foundation was laid in the 1850's by an English mathematician named George Boole. His work on what became known as Boolean Algebra was crucial to developing the digital electronics necessary to modern computers.

Much of this Boolean logic is buried deep inside the electronic recesses of your computer in the electronic circuits that make up the CPU of your machine.

However, there is an area where George Boole's work is much more accessible and just as useful. This is in the search algorithms that you can use in search engines like Google or Bing, to the search functions of your computer operating system, your department's police records management system software, GIS or crime analysis software.

Nearly everyone who has used a computer knows that you can enter a word of phrase into a Google search box to search for web pages with those words. However, most search functions also allow you to use Boolean Operators to fine tune your search. Let's look at the basics of Boolean operators.

AND Operator
The AND operator allows you to search for something that meets both criteria. For example of you search for:

ice AND cream

you will only get results that match "ice cream". You would not get results for "ice milk" or "coffee with cream".

With many search functions the AND operator is the default or implied search. For example, if you went to the search engine Bing, and searched for:

ice cream

It's the same as typing "ice AND cream". You've probably been conducting AND searches and didn't even know it.

OR Operator
Another type of Boolean operator is the OR operator. This operator allows you to search for either one, or the other. For instance if you type:

coffee OR cup

You'll get results that have either of those words in the result such as "coffee cake" and "measuring cup". OR searches tend to be much broader because the search conditions are easier to meet.

OR searches are not the default for search engines like Google or Bing so consequently you have to type the OR operator in your search terms to conduct an OR search. In the case of Bing, the OR between terms has to be capitalized. It's always a good idea to get in that habit and capitalize all Boolean operators.

NOT Operator
The NOT operator excludes a word from your search results. For example if you type:

chocolate NOT cake

then your results could return "chocolate milk" or "chocolate icing" but nothing that includes "cake".

More Google & Bing Tips
Here are a few more tips for getting the most out of Google or Bing searches. Some of these tips also have analogues for other searches within some of your applications. Check the applications' documentation for what they implement and how.

If you're looking for an exact phrase then enclose your search phrase in quotes, such as:

"the quick brown fox"

will only return results that begin with those words in that exact order. For Google searches you can even use a wildcard character, the asterisk ( * ), inside the quotes for a placeholder for unknown words. If you search for:

"the quick brown *"

It will return with results for any phrase that begins with the first three words.

You can also use two periods ( .. ) to search for a range. For example if you search for:

population 15000..25000

you'll get results that range between 15000 and 25000.

If you want to constrain your Google or Bing search to one particular website or URL preface your search criteria with "site:". In this example if you type:

burglary site:popcenter.org

it would only return results containing the word "burglary" that appeared in the Problem Oriented Policing Center's website ( pop center.org ). This is an especially handy way to narrow the results to a site that you know might be more likely to contain the information you are looking for.

Both Google and Bing have specific conventions for using Boolean  operators as well as other neat tricks. More information about search operators for Google and Bing can be found here:


Taking the time to learn how to use these operators when searching can help you find what you are looking for without returning a lot of irrelevant results.

Monday, October 3, 2016

Database Basics for Crime Analysts

I originally wrote this post for a software company’s blog in 2014. This company was bought out and recently their blog and website have been removed from the web permanently. I am reposting it here for posterity. 

One thing that is often daunting for a beginning crime analyst is understanding how databases work. Often times, just the thought of working directly with a database will cause a new analyst's eyes to glaze over. In reality though it's not that hard if you just grasp a few concepts.

So just what is a database?

A database is where data is stored for later retrieval. A recipe box full of index cards is a database. In this case, the data is instructions on how to prepare a dish such as a casserole or Granny's pecan pie.

See, nothing intimidating about this so far. Let's take this recipe box analogy a little farther.

Each recipe card is a "record". A record contains all the data needed to complete the recipe. If I write the title of the recipe on the top of each card this title is a "field". If I also have a place on the card for the number of servings, this also is a field. In this analogy we'd also have fields for ingredients, fields for the preparation instructions and field for oven temperature, etc.

Now in this example, our database is called a "flat file" database. What this means is that all the data needed for the recipe is stored in one record. Flat file databases have their use especially where the data store in each record is limited or simple. An address book application on your computer is one example.

However, if your database is going to be complex, like a police records management system (RMS) for example, it's useful to segregate data into separate tables. A table is just a group of records within a database. A relational database relates records in different tables to each other by a common field that is shared by records in different tables.

The reasons for a relational database is it is more flexible and efficient. For instance a police RMS could store information about vehicles in a vehicles table. That information could then be used by both a traffic accident module and an incident module. The data would only have to be stored once but could be used by more than one module.

Let's apply this to our recipe box analogy. If we created a separate table in our database for cooking technique, for example; grilling, frying, or poaching, then we would not need to rewrite these detailed instructions on each recipe that called for these techniques. Our recipe cards could just refer to the record in the techniques table that had those instructions.

In our recipe box analogy, a record in the recipe table could contain a field for technique and store the value of "grilling".  A record in the technique table would also have the value "grilling". These related fields are the "key". When you retrieved the record from the recipe table for grilled salmon, it would use the key field to also pull the appropriate record with the matching key from the technique table. The whole database only has to store the instructions for grilling once in the database even if a hundred different recipes called for that technique.

All this is well and good but why is it important to know how a database works?

Most crime analysts will access the information stored in their RMS though the RMS application. But there are real advantages to being able to access the data directly. For example, in the RMS I use at my department, there are several screens where various bits of data are captured and stored. However, the canned reports that come with the RMS don't display those bits when you run the report. The data is stored in the database but the report isn't configured to display it. If you can directly access the database, you can use other tools to get at this data.

When you start working with a complex database with data stored across multiple tables you need a way to know where this data is stored. Most RMS systems or other applications have a document called a "schema" that is a roadmap to your data. This schema is important if you are going to use outside tools to create reports or extract data.

At my department I have direct read only access to the database using a technology called OBDC. This allows me to access the data in our RMS using reporting tools such as Crystal Reports or to import the data into a tool like Microsoft Excel or ArcGIS for further analysis. These tools greatly expand what I can do with the information contained in my RMS database.

Now the mechanics of using OBDC to access a database directly is beyond the scope of this post. However, I can tell you it's not as difficult as you might think it is. Your database administrator can likely help teach you how to do this and walk you through some of these techniques. Once you learn how, it can really improve your ability to analyze the information contained in your RMS. It's well worth the time and effort to learn.