A friend of mine who is a Psych major relayed that there is a correlation between heat and car accidents. I was curious whether there was correlation between crime in general and weather. If you can use weather to predict crime rates, places may be able to prepare with a larger law enforcement presence when needed.
I will use the heatindex column, humid column and interpers violence column to try and see if there is correlation between the weather and crime. I will plot two visualizations to see if there seems to be a trend in either higher/lower heat index or humidity and crime.
If there is strong correlation, being able to predict days where crime will be more rampant is useful to law enforcers, and they could use this data to see when more field workers/first responders are needed.
import pandas as pd
import pandas as pd
df = pd.read_csv ('Heat_Index_Crime.csv')
df
| date | year | month | number_month | day | number_day | interpersviolence | homicides | maxtemp | humid | precipitations | windspeed | moonlight | holidays | weekends | fines | vehicles | heatindex | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1/1/2010 | 2010 | January | 1 | Friday | 1 | 31 | 2 | 33 | 78 | 0.0 | 15.4 | 1.0000 | 1 | 0 | 2 | 0 | 26.2516 |
| 1 | 1/2/2010 | 2010 | January | 1 | Saturday | 2 | 6 | 0 | 32 | 82 | 0.0 | 16.5 | 0.9870 | 0 | 1 | 13 | 0 | 30.4034 |
| 2 | 1/3/2010 | 2010 | January | 1 | Sunday | 3 | 8 | 1 | 32 | 76 | 0.0 | 14.8 | 0.9440 | 0 | 1 | 11 | 0 | 29.7769 |
| 3 | 1/4/2010 | 2010 | January | 1 | Monday | 4 | 7 | 0 | 34 | 81 | 0.0 | 10.7 | 0.8760 | 0 | 0 | 0 | 13 | 30.0638 |
| 4 | 1/5/2010 | 2010 | January | 1 | Tuesday | 5 | 3 | 2 | 35 | 83 | 0.0 | 13.5 | 0.7870 | 0 | 0 | 0 | 1 | 30.0271 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 2552 | 12/27/2016 | 2016 | December | 12 | Tuesday | 27 | 7 | 0 | 31 | 75 | 0.0 | 19.3 | 0.0602 | 0 | 0 | 0 | 60 | 30.0698 |
| 2553 | 12/28/2016 | 2016 | December | 12 | Wednesday | 28 | 5 | 3 | 31 | 79 | 0.0 | 14.1 | 0.0229 | 0 | 0 | 0 | 81 | 29.3975 |
| 2554 | 12/29/2016 | 2016 | December | 12 | Thursday | 29 | 11 | 0 | 31 | 79 | 0.0 | 23.0 | 0.0030 | 0 | 0 | 0 | 209 | 30.0530 |
| 2555 | 12/30/2016 | 2016 | December | 12 | Friday | 30 | 4 | 1 | 32 | 83 | 0.0 | 18.5 | 0.0019 | 0 | 0 | 4 | 21 | 29.3132 |
| 2556 | 12/31/2016 | 2016 | December | 12 | Saturday | 31 | 4 | 1 | 33 | 82 | 0.0 | 13.8 | 0.0203 | 0 | 1 | 18 | 0 | 32.4591 |
2557 rows × 18 columns
Obviously, other factors play an effect on crime. For example, there might be more violence on January 1st as this is New Years, and typically a holiday where many are under the influence. The table does show if a day is a holiday, which could help explain some outliers, if they occur, but other instances not shown, like maybe a protest, could lead to unexplained outliers.