Let's look at murders in Philly

[malbasi]

Please complete all of the following sections, or the ghost of Joseph Pulitzer will spookily dance around your issue! A completed version of this template can be found at https://github.com/jsoma/data-studio-projects/issues/1

Pitch

Summary

I looked at data the Washington Post collected about murders in cities across the US. I analyzed that data for Philadelphia.

My data came from WaPo's GitHub

Details

Possible headline(s): How to get Murdered in Philly

Data set(s): https://github.com/washingtonpost/data-homicides

Code repository: https://github.com/malbasi/data-studio/tree/master/code/01-murder

Possible problems/fears/questions: I'm not sure the best way to integrate city-wide demographic data with the data I have.

Work so far

Checklist

This checklist must be completed before you submit your draft.

• [x] I have already spent time with my data set, opening it, exploring it, etc
• [ ] I have created a "DIARY.md" file to save links and list all of the terrible, no good problems I come across
• [ ] My issue issue has [Project] in the title
• [x] My issue links to my data set(s)
• [x] My issue links to my code repository
• [ ] My issue explains what I'd like to explore in the data set
• [ ] My issue includes images - either inspiration or what I've done so far
• [ ] I have received two comments of peer feedback on my pitch issue

[vpenney]

• I like the different ways that you visualize data, between bar charts and the scatter plots. It helps paint a picture for the reader.

• It would be interesting to look at the data year-over-year, or to find a new way to categorize the data on your scatterplot (maybe a choropleth? by what, I'm not sure) just to see if there are any correlations.

Nice work on styling your charts and graphs!

[castorsia]

Well-targeted set of questions that all lead to visuals. Loved the color gradients. Maybe breaking down Police station arrests would be an idea?

[nickospi]

• Really nice charts / like the pale white/yellowish background / like the fact that there's a variety of charts
• Would add different colors in the race scatterplot

[malbasi]

Update

I made a few changes to the scatter plot (added legend, coded by race) and cleaned up portions of the other charts.

Your project content: images/words/etc

Any changes in direction or topic?

Direction and topic stayed the same

Problems/Questions

Making a legend in matplotlib is absolutely awful. I eventually gave up and completed the work in illustrator instead. I also really wanted to make a plot of how murder changed by race over time (like the first graph, but with different lines for different victim's races), but was hampered by the structure of the data.

Checklist

• [x] I have included my visuals
• [x] I have filled out the sections above
• [] I have been updating my DIARY.md with details about my process
• [x] I have uploaded/updated any Jupyter Notebooks or other datasets into my code repository

[angelareplica]

Really great and thorough! Impressed by the matplotlib work. For the stacked bars, I think it might be visually more clear if they were stacked in order of quantity, rather than occasionally having the thinnest sections in the middle or at the beginning. (Hopefully that made sense.) If it's in the data, it would also be interesting to see the racial breakdown of the arrests/perpetrators. Maybe a little more context for the scatterplot too: Are murders concentrated in racially segregated neighborhoods?

[zle2105]

Great job! For the Murders by Race chart I would maybe change the middle color value to something brighter. For the Murders Per Year in Phili line chart, it may look neater if the blue is not filled in to the bottom. Overall, I really liked the style and the scatter plot map.

[jsoma]

• I think fat bars take up too much space, even though making the graphic bigger makes it seem like one has accomplished something! Thin them out a bit, just by changing the aspect ratio of the graphic.
• I would probably ditch the horizontal lines, only keep the grid lines that work with the axes to tell you numbers. You don't need one coming out of the middle of a bar, y'know?
• I'd add some spacing around the graphic, just so it's framed a little more gently
• I think the question here becomes, "what am I looking at?", or "if this should be interested, what should I be comparing it to?" - for example, you have the arrest records v. different victim races. Mostly this tells me most of the times there's an arrest, sometimes it gets closed without anything. But if I'm supposed to be comparing, we'll want to make it percentages instead of actual numbers. The actual numbers are so disparate between black and asian that I can't make any sort of comparison, I'm just comparing between the bars on 'black', but those bars don't mean much because they're the same general order for each race. The big thing that jumps out is white murders have a LOT higher % of arrests than anything else, but it gets kind of lost in everything else that's going on. So basically what I'm saying is, make a chart that shows that percentagewise, white victims get their cases solved by arrests all of the time.
• If you do go with the many charts at the bottom, my first instinct is "too many labels" or "too many ticks," but you did an excellent job keeping them with the same x range, so the problem would be if you only have 500's then the asian one would be more difficult to read, etc. No real answer off the top of my head, but if you think of something feel free to give it a shot.

[playfairbot]

Hello! I'm a little robot, here for a surprise inspection.

You need some feedback, let me summon @Yuanqi-Hong, @linleysanders, @adrianblanco for you

[malbasi]

Final

Project visuals/text

Since 2007, murder rates have been in decline.

However, the rate of murder for certain races continues to be unequal.

And the people who get justice for the murders of their loved ones are not those who are the most victimized.

Details

Headline: Murder Is in Decline, but Is It Fair?

Code repository: https://github.com/malbasi/data-studio/tree/master/code/01-murder Final data set(s): https://github.com/washingtonpost/data-homicides

What did you find to be the most difficult part of this project?

The wrangling was probably the most difficult part of this project. While the data was very clean, it required some re-shaping to be useful for my purposes.

Are you satisfied with what you produced? Is there anything you would like to change or improve?

More or less. I think the graphs could still be tuned in a little and their style could be tightened up. I would have liked to include a map component as well to show the geographic distribution of the murders too.

Checklist

• [ ] I have included my visuals
• [ ] I have posted my project to my project website
• [ ] I have been updating my DIARY.md with details about my process
• [ ] I have uploaded/updated any Jupyter Notebooks or other datasets into my code repository"

[Yuanqi-Hong]

• In my opinion, murders per year should be a bar graph rather than a line graph. I think the data here points to the actual, absolute number of murder cases each year? In that case it's not really a "rate" that fluctuates over a consecutive time period. That said, maybe consider dividing the murder (deaths?) each year by the population? (Bear with me I'm a big believer in normalized data) Like, if there're 2 people living in Philly then the worst that could happen would be they killed each other and the number of murders for that year would be 2. The absolute values will make sense, however, if Philadelphia's population doesn't change that much at all from 2007 to 2016.
• I like how you put shades on that last graph. It separates the groups clearly. (The white background is invading Asian and Black parts, though. Maybe this looks better when the background colors around those two groups and of the page are different.)

[adrianblanco]

That said, maybe consider dividing the murder (deaths?) each year by the population? (Bear with me I'm a big believer in normalized data) Like, if there're 2 people living in Philly then the worst that could happen would be they killed each other and the number of murders for that year would be 2. The absolute values will make sense, however, if Philadelphia's population doesn't change that much at all from 2007 to 2016.

I guess the population did not change so much. But stealing this suggestion I'd recommend to normalize the data too. Then you will be able to compare Philly with to other big cities in the States. Them, the context of the story will be richer and you'll be able to point if this is a big problem in Philly compared with... NYC, Chicago or whatever.

[sarahslo]

murders per year should be a bar graph. it's a discreet number, fever lines are used when you have an average, an index, analysis or if you need to plot several trends.

what order are these data in? they are not in the order of the data, they are not alphabetical. they need to be arranged with some logic.

also with the keys and the colors, the key should be in the same order as the data. if closed by arrest is on the bottom of the chart then it should be on the bottom of the key so our eyes don't have to go all over the place looking. as to the color, i'd make them more distinct from each other. they are all the same tone. they need to separate more so it's hard to see the difference.

except in the percentages one (again, what order?) but here, you need to be careful because the numbers are so different -- there are hardly any murders of asian people, i don't know that the percentages are significant.

what order are the charts in here? should they tell a story? is there a point you are trying to make here with the charts? is there some surprise i should take away or some way i should feel informed? i don't know what i'm supposed to be focusing on here in terms of the data story. if you figure out what you want to say that will inform the order you put the data in, how you use the data, what order the charts are in....

[malbasi]

It looks like Wa Po did a similar story looking at the racial breakdown of murders. Their charts are beautiful https://www.washingtonpost.com/graphics/2018/investigations/black-homicides-arrests/?noredirect=on&utm_term=.1ab04847dad0