data-science
data-science copied to clipboard
hackforla
CoP: Data Science: Identify overall trends in DUI arrests by council region
Dependency
~Access to clean 311 data from 311-data-org team~ This product should NOT be correlated with 311 data Project one sheet (can continue issue w/o completion of one sheet template)
Overview
Investigate areas of high and low documented DUI arrests. Link this data back to 311 data to provide insight for councils.
Action Items
- [ ] Revise the overview to remove correlation with NCs
- [ ] Find available data sources and add to Resources section below
- [ ] Determine is this is one-time or ongoing project (and assign appropriate label)
- [ ] Write one-sheet
- [ ] Define stakeholder
- [ ] Summarize project including value add - needs to be informed by a ux interview with some NC leaders
- [ ] Define project 6 month roadmap
- [ ] Detail history (if any)
- [ ] Define tools to be used to visualize combined data
- [ ] Create issues for the following
- [ ] EDA (Exploratory Data Analysis) of DUI data
- [ ] Investigate areas of high and low DUI citations
- [ ] Investigate correlation of DUI with population and percentage of stops
- [ ] Investigate correlation between DUI and collision data
Resources/Instructions
Sofia says, in order to determine if its a one time project, that we should see how much data there is.
Resume Skill
Perform EDA and independent analysis into public safety data stakeholders are nc who might do outreach campaings (e.g., offering uber vouchers to bar tender to give patrons, etc).
Have completed initial analysis comparing rate of DUI's in each neighborhood district over time as well as areas most impacted during an overall drop from 2013 to 2016. Need to identify source of population data per neighborhood district.
Identified dataset for population by NC District. Added to Resources Markdown document.
There are outliers, but the overall trend seems to associate presence of Liquor Stores with DUI's. This is looking only at 2019 arrest rate since the business data is not historic.
There is some correlation between population density and DUI's but it doesn't account for the neighborhoods with the highest numbers.
I need to investigate the poverty numbers that came from the city data portal. It seems unlikely that the neighborhood with the lowest poverty level still has 75% of residents below the poverty line.
Pulled raw data from Census.gov site and ran calculation of percent below poverty line by neighborhood. The only unexpected outlier was the neighborhood with over 50% below poverty line. After investigation, the neighborhood boundaries line up with the UCLA campus boundaries, so the numbers are being influenced by students without income.
It looks like you're continuing on this. Have you found anything you feel like we should incorporate into the site?
@johnr54321 There are definitely neighborhoods where DUI arrests are more prevalent. If there is interest from the Councils, I could see it being integrated either as a point layer or heat map. I know the city also has a campaign meant to reduce pedestrian fatalities. That was another dataset that could be of interest to the target audience.
@JRHutson Do you have something visual (like maps, pins, or chart) on the DUI's? Also, was wondering if you had access to a dataset on cellphone movement.
@johnr54321 I'll send you some examples via Slack. I'm assuming you're looking for general examples to get feedback from the NC contacts on what they would find useful.
I have not seen any datasets on CellPhone movement publicly available.
@akhaleghi what I had been working on is in a repo under my account. Happy to fork it to the Hack For LA account incase the code could be useful to anyone. I got it to a point that there was some insight but not a finalized presentation. https://github.com/JRHutson/dui_analysis
@ryanmswan should Regan commit this to the DS repo or do you want these files somewhere else (i.e. a separate repo)? @JRHutson thanks for this! is this project you want to continue working on? if so, should we get someone else to help you or, if not, do you want to hand this off to someone else?
Abe,
I think that handing it off would be ideal if there is still interest in the data from the Neighborhood Councils.
I’ve been helping the Home Unite Us team as a PM and trying to get out of a toxic team at work, so I don’t expect to have bandwidth for it any time soon.
Regan
Sent from my iPhone
On Oct 8, 2021, at 1:30 PM, Abe @.***> wrote:
@ryanmswan should Regan commit this to the DS repo or do you want these files somewhere else (i.e. a separate repo)? @JRHutson thanks for this! is this project you want to continue working on? if so, should we get someone else to help you or, if not, do you want to hand this off to someone else?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.
@JRHutson Thanks! If you could fork it to the Hack for LA account, that would be awesome.
current status of this issue is
- Someone needs to talk to Regan to get him to commit the data to the data science repo
- Note where he has put it in the top of this issue (for easy access)
- Review issue again for next steps
@JRHutson does it make sense to put in into a folder on the ds repo called dui_analysis? If so, please commit it there.
@ExperimentsInHonesty Looks like I need elevated permissions to fork it or someone else needs to do it. Let me know either way.
Connected with Bonnie last night and was able to fork the repo to https://github.com/hackforla/dui_analysis. @akhaleghi let me know if there is anything else you need me to do on my end.
@salice @KarinaLopez19 Bonnie wanted us to look at this to determine if there is more that should be done on this issue. Any thoughts? If it's completed, maybe we can have a volunteer create a presentation.
Next steps: Review the data output above and create a write-up and presentation with your own analysis.
None of the top comment are checked off. This issue needs a whole re-write so that someone could actually pick it up. I am happy to give a clinic in how to do that to @akhaleghi and @TaniaKhan04 . Please make an appointment with me.
@loboramiro are you clear on what you need to do on this issue, or do you need us to do a rewrite first?
