Criminal_Weather_Denver

Project Title: Criminal Weather

Team Members: (git repository user name)

• Melanie Nolker (mnolker)
• Lisa Stroh (Lisb020)
• Kathy Nguyen (kcknguyen)
• April Lagnevall (alagnevall)
• Haifa Najdawi (HaifaNajdawi)

Project Description/Outline:

• Does weather have an impact on a crime? Find if there is a correlation in non-traffic crime in the city of Denver compared to various weather conditions.

• Hypothesis: Weather has an impact on non-traffic crime in Denver.
• Null Hypothesis: Weather does not effect non-traffic crime in Denver.

Research questions to Answer:

• Is there a correlation between crime & weather?

• What type of offenses will be used to analyze crime rates?
• What weather conditions will be used to compare to offenses?

Datasets Used:

• Violent Crime: https://www.denvergov.org/opendata/dataset/city-and-county-of-denver-crime

• Denver city crime from 1-1-2016 to 9-9-20.csv
• filter for "Is Crime" = 1
• excluded traffic related offenses
• Grouped .csv data by:

• Occurrence_date
• OFFENSE_TYPE_ID
• OFFENSE_CATEGORY_ID

• Weather API = https://www.worldweatheronline.com/developer/api/docs/historical-weather-api.aspx#qparameter

• API and pulls weather data for every day from 1/1/2016 to 8/31/2020
• pull muliple weather conditions

• avg temperature (F)
• avg precipitation (inches)
• avg cloud cover (%)
• avg wind speed (mph)

Anaysis Performed:

Seaonality:

Scatter plot of Number of criminal offenses & Weather measures (y-axes) over 5 years (x-axis):

To further explore the seasonality of crime data in Denver: Grouped 5 years data by months to view in boxplots the distribution of the data’s yearly seasonality

Split weather conditions & view breakdown of criminal offenses of those split conditions:

Bar chart of Offense type Split by Temp Mean: (see Jypyter notebook for split of other weather conditions)

Results of Analysis

Scatter plot of number of criminal offense compared to each weather condition:

Statistical results of coorelation:

Based on the p-value of # offenses to each weather condition…

Reject Null Hypothesis:

1. Avg Temp (F) = 9.06 e-42
2. Cloud Cover (%) = 2.23 e-06
3. Precipitation (in) = 1.02 e-13

Failed to Reject Null Hypothesis:

1. Wind Speed (mph) = 0.89

The results of the correlating (r-value) # offenses to each weather condition indicate a weak relationship between crime and weather conditions:

1. Avg Temp (F) = 0.32
2. Cloud Cover (%) = -0.18
3. Precipitation (in) = -0.11
4. Wind Speed (mph) = -0.0

Rough breakdown of Tasks by group members:

• Project coordinator / merger - Melanie
• Crime dataset anaysis:

• Kathy
• April

• Weather dataset analysis:

• Haifa
• Lisa