Ten ways to make your crime maps more ‘interesting’

PFPE mapAt various conferences and visits to police stations I have seen quite a few maps, and some have been great – really well presented, laid out and prepared. Alas many are awful, so I’ve put this page together as a brief guide to those perhaps less versed in the cartographic ways. You do not have to adhere to the guidelines here, but they might improve the readability and quality of your maps. If you do not follow these suggestions, ArcGIS will not self-destruct in a fit of cartographic rage – but this is part of the overall problem. The software does not understand your data and will let you do just about anything you want – even if it is wrong.

Note: If you are a Computer Science major and confused by the concept of sarcasm, feel free to click over to the vanilla version for you earnest types.

Tip 1: Do not include a scale bar. This will make it much more interesting as your map readers have to guess the distance between objects. One of the main aims of mapping crime is to compare areas and examine the proximity of objects, so why make it easy for the uninitiated to understand you map? Without a scale bar nobody will have a clue how far things are apart and this gives you the opportunity to have impromptu quizzes or make things up as you are presenting. If you accidentally include a scale bar use a scale that goes “0 —– 6.75 ——13.25 kilometers” instead of the usual “0 —5 —10” or similar. Big complex numbers really impress audiences.

Tip 2: Do not include a North arrow. Hundreds of years of cartographic tradition have no place in the new millennium – we are in the digital age and therefore all maps automatically have North at the top: even if we have to rotate the map to get it to fit on the page. Anyway, if you have visitors from outside your suburb, city, or country, why should they want to know in which direction is North so they can orientate themselves? They probably are not interested anyway.

Tip 3: Use jargon and special codes in the title of the map. Including special codes and police service jargon in the title of your map will make it, and you, look more professional. A title such as “B-type crimes for sectors GF and YTU for shifts R4 and R5” really impresses audiences. Make sure you also use dates in a mixture of European and American format at international conferences (without telling the audience which you are using). After all, 10/10/00 is the same either way, and what can the rest of the world teach us? Better still, don’t have a title at all (or have one that warbles on for three or more complete lines), and never put your name on the map – that way there is nobody to blame.

Tip 4: Find the color palette, and use every one. Color is what maps are all about. Use as many colors as you can find. There really are no rules about inappropriate choices of color, so bright cheerful pinks are fine for displaying child murder sites. If you have interesting symbols at particular places (such as body dump locations) try to de-emphasis them by making the background color glaring and bright. This will detract from your murder sites and make the viewer only see the underlying light industrial land use – much more important. Other features such as roads and railways and national parks are probably not very relevant, but they fill up the map nicely so give them a bold, bright color. This will further detract from your important points and distract the viewer – making them think there is less crime.

Tip 5: Don’t worry too much about understanding your data. The important thing is the presentation and the display. Don’t worry about showing maps with dots all over the place, often obscuring other dots. The audience will get the general picture and do not need complicated things like graduated circles to show how many crimes have occurred at the same place. This is just pedantic mapping for airy-fairy academics. And certainly do not worry about it when showing maps of repeat victimization. Also, if you can only geocode to the level of a zip code, still show the viewers the very best detail you can, right down to the street corner. Let “spurious accuracy” be your guide.

Tip 6: Make the most of thematic mapping. Those automatic thematic map menu items are there to be used as much as possible, and negate the need to really understand what they do. After all, it always looks great so it must be right! If you have a numerical date variable, use the graduated circle. I especially like those maps that show the time of day of an offence as a graduated circle. The bigger the circle – the later in the day. Stellar cartography right there. Another one is to use the bar graph function when comparing big numbers and little numbers. You can never see the little bars unless your nose is against the screen – how that one makes us all laugh in my office.

Tip 7: Legends have had their day. In bygone years there was a time for legends, but that age has most definitely passed. In modern cartography – especially for presentations, you will be there to explain the symbols and the values associated with different colors. And if you forget, don’t worry – the audience will understand. Hey, we’ve all done it. If you have to be passé and include a legend, then for color shaded areas use impressive looking numbers such as “3.01453 to less than 6.03215”, instead of “3 to less than 6”. This will impress the audience no end as you obviously have a grasp of quantum arithmetic.

Tip 8: Caveats look weak. You are out there to impress with your map. Having a caveat, especially for the geocoding rate, looks weak and as if you have not put enough effort in. To suggest that you have not been able to map every point will make you look bad next to all the other crime mappers who must obviously be better at it than you. After all, how can you make a good impression if you have data error? Being misleading is just helpful.

Tip 9: Get as much information onto a map as possible. Maps take time to produce so it is important to squeeze in as much information as possible. This is especially true for maps with symbols. Try and use more than 5 different types of symbol on a map, and ideally make them roughly the same size and color. It would be unfair to give added weight to one, so make them as indistinguishable as possible. If you can make them illegible from the back of a room on a PowerPoint presentation then that also helps because it makes people have to concentrate and come closer.

You could also try to disguise unhappy symbols like the locations of assaults and murders with unrelated symbols such as public lavatories and libraries. This can of course work both ways, it might suggest that there have been a lot of robberies, but most of your viewers will be fooled into thinking your area is well stocked for public utilities.

Tip 10: Never let anyone photocopy your map. A map is a work of art and should never be disseminated – ever. Stick it on the wall in the office, and use it in presentations but never let it out of your sight. Someone might take it down from the wall and photocopy it – ruining the point of the thing. Worse, they might actually use it to make a good public safety decision. The best way to teach them not to use your work is to make symbols and background colors roughly the same shade. A mid-blue and a mid-red blend nicely in color, producing a pleasing effect on the eye, but are indistinguishable when photocopied into grayscale. That will teach them to steal your maps!

Jerry’s top ten PowerPoint tips

After all the work that some people put into their research and analysis, it sometimes defies belief that they make a complete hash of their moment to impress – their moment to shine as an intellectual. After months of slaving over a hot keyboard many people use the opportunity presented by a briefing or conference to confuse the audience, contradict themselves and generally disappoint the whole room.

They do this by having a lousy PowerPoint. Now I’m well aware that some people reading this will have seen my presentations, so I’m not claiming to be a great presenter. But I would like to think that my audience at least can see and understand my slides, have time to read the words, and understand what I’m on about (most of the time). Having spent days or even weeks of my life staring at innumerable incomprehensible PowerPoint presentation, here are ten tips that I think help avoid some basic traps.

A PowerPoint presentation in the briefing room at Philadelphia Police Headquarters
A PowerPoint presentation in the briefing room at Philadelphia Police Headquarters

Tip 1: Strike a sensible contrast between text and background. Right at the beginning, try to get the presentation off to a good start. Back a few years ago when projectors were in their infancy, PowerPoint presentations looked best with a dark background and light text as these were the easiest to read and most gentle on the eye. That rule is still true, and by using a dark background you can make better use of a range of colors, colors that often look faded or indistinguishable against a light background.

The one caveat with dark backgrounds is that in brightly lit rooms, they can look washed out. So an option is to switch to a light background with dark text. Modern projectors are now able to project this combination better than before. But be warned that in a dark room with a bright PowerPoint presentation, the people near the front, stuck immediately in front of a huge, bright, glaring screen will feel like they are being interrogated. Be gentle on the audience and they will appreciate your message all the more.

Tip 2: Use simple titles and points. Long titles that go on for more than one line is a no-no. Keep titles and bullet points short and relevant. You are the message, not your slides, so don’t overcomplicate them. If you are a more accomplished speaker and use the slides as a prompt (always better than reading a script) try a few quirky bullet points. They will act as a more memorable prompt for you and will intrigue the audience.

Tip 3: Get the font size and type right. Absolute minimum font size 16, but bear in mind that with some fonts (such as Garamond) 16 is smaller than in another font (like Arial or Calibri). Flowery, airy-fairy fonts do not work well as they are difficult to read. Simple San Serif-type fonts such as Arial are simplest to read, but always go to the back of the room before the presentation to check the legibility. If you can not fit your text on the screen using a font size of 16 or more, don’t reduce the font size – reduce the number of words.

If you are taking your presentation somewhere else and using another machine, do not use some unique font you downloaded from the net, such as Ratcliffe’s Bizarro Bold Font (not a real font, I hope). The host presentation machine is unlikely to have the font. Unless you are skilled enough to have embedded the font in the presentation, PowerPoint will default to a standard font and this will probably disrupt your layout. Stick with the basic fonts unless you know what you are doing with embedding.

Tip 4: Limit the number of bullet points. Never more than 7 per slide and about 5 is best (think 4-6 as a good rule of thumb). A presentation should be an illuminating summary of your work, not the whole damn thing – so summarize. If you really want to put in more bullet points then break your list into a number of slides with other stuff in between. Listening to a presenter reading a list that can be read faster on the screen is many people’s idea of hell. Instead of boring people with a long list, give them a handout. Better to tell them a few things well, than lots of things badly.

Tip 5: Avoid the trap of fancy builds and dimming. ‘Builds’ is the term given to those fancy swirly ways to introduce text or other items to the screen. I’m sure you’ve seen them: text flies in from the left, then one letter is added at a time from the right, etc. If you need to use fancy builds to impress your audience then you really are struggling… Whizzy builds tend to annoy audiences, especially the slow builds accompanied by applause sounds or camera clicks. Regular conference attendees have seen them all and are not impressed. They also distract from what you are saying and your message. Slow builds can also be a presenter’s nightmare because if you have to hurry through your last slides, they hold you up. Avoid slow or complicated builds and transitions between slides. Use only the simple stuff.

Dimming is the term given to the dulling of, or worse, disappearing text or objects once the next item is visible. Legislation should be in place to prevent presenters showing you a bullet point and then hiding that point when the next point arrives. Most conference attendees have a modest attention span (and I’m being generous here). It is therefore annoying for them to look up after a few seconds mental time-out (i.e. checking their iPhone) and find they have missed the first points. Never dim segments of a chart as chart segments are only valuable if seen in proportion to the other elements.

Tip 6: Don’t rely on the spel chequer. This should be an obvious one, but I’ve seen one academic talk to a room full of 600 law enforcement personnel with slides that said pubic order instead of public order. Form instead of from, policy instead of police – there are lots of common mistakes. The importance of maintaining pubic order is still a favorite of mine though… Remember to plan your presentation prep well in advance: the more you are in a rush the more you will make mistakes here.

Tip 7: How to bore – include technical detail. There is nothing more soporific than large equations on a PowerPoint presentation (unless your whole presentation is to a room full of statisticians about a new algorithm you have just discovered). Complex, illegible flow diagrams with too little time spent explaining them can also do it. The presentation should be a short, catchy and impressive summary of your work. Don’t impress them with how much you have done, just impress them with what you have done. If you must include technical detail, give them a handout afterwards, or better still, point them to your most recent book/article/publication.

Tip 8: Maps and graphs speak volumes. To help with maps look at my top ten crime mapping tips and I suggest you look at this for charts as well. Maps are great for PowerPoint presentations as a picture really does say a thousand words. Never assume that your audience will know where you are talking about (for example, entertainment can be had asking Americans to name the capital of Australia or Canada). Charts and graphs should be simple and never have more than about 5 pieces or components. Line and bars are good for showing time periods, pies must show parts of a whole (i.e. 100%) and surface charts can show general trends. Don’t cram too much in, and don’t dim any part of a chart (see tip 5).

Tip 9: Continuity across slides is essential. If you really want to annoy your audience, change the builds, fonts, colors and styles regularly throughout your presentation. That will really hack them off, and it looks really amateur. Good presentations are slick, professional and keep up a consistent style throughout. Your organization may already have a corporate style, but don’t feel you have to throw away the tips here if they do. Decide on a color scheme, set of one or two fonts (One for titles, one for text) and stick to them.

Tip 10: Finish on your title slide or a black screen. You just finished your presentation, it has gone superbly well, and then you give away all the magic and let the audience see the trap door underneath the stage (i.e. the slide sorter page). When you finish talking show your title slide again with your contact details or end on a meaningful quote so that they can concentrate on your message and how good you are. Don’t let them see the slide layout view at the end, especially if you have unused or hidden slides. Doesn’t look good and will distract the audience just when you had won them over.

What we have learned from Philadelphia foot patrols

With the recent publication of our comparison of foot patrol versus car patrol, it ‘s worth a quick review of all that we learned from the Philadelphia Foot Patrol Experiment. Especially as the paper Liz Groff took the lead on is available for free from the publishers until the end of May.

The original foot patrol experiment paper described our randomized controlled field experiment which saw the Philadelphia Police Department place 240 officers on 60 violent crime hotspots (randomly selected from a list of 120) for the long hot summer of 2009. And it was hot walking the streets of Philadelphia in a ballistic vest – we all remember the fieldwork and empathizing with the officers who did it all summer!

At the end of the experiment, 90 violent crimes had been prevented, resulting in a net reduction of 53 violent offenses after some displacement. This was a 23 percent reduction in violent crime as a result of foot patrol in carefully-targeted areas – a unique finding for policing.

How was this achieved? We found that pedestrian stops increased by 64 percent in the foot patrol areas, probably increasing the likelihood that offenders would be stopped, and subsequently reducing their enthusiasm for carrying a firearm. We learned some other things that summer:

  1. There was no community backlash within the foot patrol areas. To the contrary, members of the local community were really upset when their foot patrol officers were eventually removed, and they let the PPD know about it in no uncertain terms.
  2. The image of foot patrol as a punishment posting changed to a degree within the PPD. Good commanders became convinced that foot patrol was a practical tactic in high crime areas, and some patrols remained in place after the experiment.
  3. The fool patrol officers got a real feel for their foot patrol areas, developing community and criminal intelligence in the months they spent on foot.
  4. The foot patrol officers engaged in more pedestrian stops than their vehicle-bound colleagues, and they also dealt with many more disorder incidents – an activity that is always an issue in the summer in Philadelphia. They dealt with fewer serious crime incidents, yet were undoubtedly responsible for the decline in violent crime.
  5. Importantly, the foot patrol officers engaged is a different type of police work than their colleagues in cars. Less response-driven, they engaged in more order maintenance and community-related activities. They did not replace the activities of the cars, but rather work in a complementary fashion, being co-producers of community safety with their colleagues. Even if they sometimes wandered a little.

Unfortunately, we also learned – in a subsequent Criminology article headed up by two enterprising graduate students, Evan Sorg and Cory Haberman – that the gains achieved during the foot patrol experiment did not last. The effects dissipated as soon as the foot patrol officers were removed, and in fact some effects were starting to wear off as the foot patrol experiment continued into the late summer.

My colleague Jen Wood took the lead on the qualitative component so important to understanding the nuance of the foot patrol experiment. We learned that officers negotiated order based on geography, people and space, and varied their strategies and tactics based on their knowledge of the people and the environment.

The experiment was generously awarded with a research award from the IACP and from the American Society of Criminology’s Division of Experimental Criminology, but more importantly it helped people recognize the Philadelphia Police Department as an innovative department willing to try new things, take risks, and learn. And while it involved a lot of researchers, they nearly all volunteered their time on top of their normal duties. Temple University and the College of Liberal Arts generously helped out with some fieldwork costs, indicative of their desire and ongoing commitment to moving the city forward; But the experiment – which learned so much – did not cost the city taxpayers a single cent.

Why we should address risky places

One of the benefits of the move to data-driven policing is the opportunity to understand how a few risky places disproportionately contribute to the crime problem. Research conducted here in Philadelphia has already identified that violence appears to strongly cluster within 85 feet of bars and extend for up to a block, robberies concentrate around subway stations, and drug arrests can cluster around pawnshops and check-cashing stores. In the last month or two new research by Elizabeth Groff and Brian Lockwood and based on Philadelphia Police data finds that even when controlling for the population characteristics of the immediate area, violent crime is nearly 10% higher on blocks with bars, and disorder is more prevalent up to three blocks from subway stations and schools.

We also know that roughly 20% of places are responsible for causing 80% of the problems for police. Crime scientists call this the Iron Law of Troublesome Places. The iron law holds for banks, convenience stores, gas stations, schools and a whole host of other places. It also holds for bars and subway stations, so it is likely that only a few bars in the city are responsible for the vast majority of major crime around all Philadelphia bars. Now that every district has an analysis coordinator to help with identifying these key locations, the next step is to engage L & I to close the worst places down, or to find other ways to “encourage” owners of the most problematic bars to be more responsive to the crime problem in and around their facilities. In this way the police department can make best use of available resources and engage other city agencies and business owners in the fight against crime.

(A version of this post was first printed in the Philadelphia Police Department ‘Call of Duty’ newsletter under the Intel Driven column).