Embedded Files
This project aims to gain insight into the influence of United States mental health care shortages on local crime rates. It uses two sources of data. 1. The US Health Resources & Services Administration (see data.hrsa.gov/topics/health-workforce/shortage-areas ) and 2. the FBI Crime Data Explorer (cde.ucr.cjis.gov/LATEST/webapp/#/pages/docApi ).
US health provider shortage areas, specifically mental health are more conveniently available to the general public in the form of an interactive webpage dashboard (see left image) but i have extracted from bigquery-public-data.sdoh_hrsa_shortage_areas.hpsa_mental_health
The FBI Crime Data API is a read-only web service that returns JSON or CSV data. It is organized around the data reporting systems the FBI UCR program uses and their related entities. Agencies submit data using either the Summary Reporting System (SRS) or the National Incident Based Reporting System (NIBRS).
For the Jupyter notebook and more info please go to github.com/echanj/MHSA_vs_Crime_geomap.git
(left) MHSA locations short listed for this stage of the project and mapped using folium. The heat-mapping is based on the HPSA_Score : Health Professional Shortage Area (HPSA) Score developed by the National Health Service Corps (NHSC) in determining priorities for assignment of clinicians. The scores range from 0 to 26 where the higher the score, the greater the priority.
(left) The NIBRS agencies are identified by enumerating through all above MHSA locations and then using help of a Haversine function to determine which agencies are within a 20 mile radius of any MHSA. then for each MHSA we can now access specific crime data such as date and number of specific criminal acts (eg. rape, murder or traffic violations)
(below) Selected EDA and results from the project (for details visit github.com/echanj/MHSA_vs_Crime_geomap/blob/main/mini_project_main_2023.ipynb )
For the Hypothesis testing 𝐻0: "That there is no difference in means for an NIBRS agency crime data near an MHSA location before and after the date the MHSA status had been withdrawn."
This was tested for MHSAs in each state. For the majority of states, the alternative hypothesis is rejected. i.e., for each state The total percentage of t-tests that failed to reject the null hypothesis was below 50%. This suggests that there was no relationship between the improvement in the availability of mental health care providers in the shortage areas and the incidence of crime near those areas.
Indirectly, the results suggest that crime is not as influenced by the state of a person's mental health as we may be led to believe. It is more commonly assumed that criminal acts are performed by people whose mental health is degraded and requires diagnosis and treatment. This may not actually be the case.
We might have tried to identify a relation between the MHSA parameters 'Degree_of_Shortage' or 'HPSA_score' and crime rate; however, the NIBRS data does not have any population information so we cannot normalize the crime counts between agency locations. For future reference, it should be possible to obtain this information by other means of another API or even restrict the original HSPA search to entries which contain some population data.
Link: notebook page
Link: html version
Page updated
Google Sites
Report abuse