Where are CCDAV's frequent callers and what do they want?

by Archana Adhikari and Azka Hassan

Last updated on 1/18/2022

for Data Visualization for All
with Prof. Jack Dougherty
Trinity College, Hartford CT, USA

Introduction

Seeking to support domestic violence victims, CCADV (Connecticut Coalition Against Domestic Violence) often interacts with callers from all over the state and provides a wide array of services. For many cases, 2 or 3 interactions are simply not enough, nor is the provision of 2 or 3 services. In order to identify patterns in regions among frequent callers and the services that pertain to those calls, we are trying to answer the following question; Which services are recurring amongst frequent callers in each region?

We believe that CCADV can greatly benefit from learning more about their frequent callers, where they tend to be located, and the services that they receive. It can provide valuable insight on what services are effectively helping their callers to resolve their problems. The frequency of calls before the resolution of the problems may also reveal the effectiveness of the services provided to them. Being aware of their callers, their location, and the services they are likely to receive can help CCADV be well-equipped in helping their frequent callers based on their location.

Data

We began by defining some of the terms that will be used in our data. Frequent callers are those who called CCADV more than 11 times. We calculated the median number of calls among all the callers from 2020 to 2021 excluding 1-time calls so that we can avoid the inclusion of outliers such as 75 calls affecting our data.

We then made pivot tables that summarize calls greater than 11 in each region which includes CCADV’s services. The services provided by CCADV are the following: Counseling, Victim advocacy, Civil Legal, Criminal Justice, Safety planning, Info provided, and Referrals. We found that 100% of the frequent callers have used at least one of these services. We normalized our data by calculating the percentages of different services provided by CCADV in each region for frequent callers. We also calculated the mean percentage of different services provided in all regions of Connecticut, and in every region, we selected those services that are above the mean percentage. For example, we calculated the percentage of calls that include counseling services, calculated a mean of those percentages, and selected those regions in which counseling percentages were higher than its mean. This would mean (pun intended) that the frequent callers from those selected regions are more likely to utilize counseling services than the frequent callers from regions where counseling percentages are below their mean.

The following is a map that shows the percentage of callers in each region who reached out to CCADV more than 11 times from 2020 to 2021. We can see that according to the CCADV data, the New Haven region has the highest percentage of frequent callers, followed by the Hartford region, and the New Britain region. All these callers have used at least one of CCADV’s services.

This table consists of the percentages of different services used by frequent callers. It shows how Counseling and Info Provided are two of the highest demanded services among all services used by CCADV’s frequent callers followed by Victim Advocacy.

The map below shows the most prominent services provided in each region of Connecticut by CCADV along with the percentage of that service used in the region. This illustrates how counseling services dominate most of the regions except the Mansfield region. Information service is popular among the frequent callers in the Mansfield region.

The CCADV data defines counseling services as "anytime an advocate spends more than 10 minutes talking to a client." With these findings, CCADV may consider allocating resources to train staff that specifically provide counseling services. In one of our meetings, a CCADV staff member, Rosario, mentioned that there is a distinction between the type of counseling CCADV data refers to and clinical psychological services. She also expressed a need for CCADV to provide those clinical counseling and psychological services. With our map showing such a high demand for counseling services, we believe it will help CCADV in their future decisions in refining and improving the counseling services they are offering.

The chart below shows the services in each region with a higher percentage than its mean percentage of service:

This chart illustrates how overall, CCADV heavily provides Counseling and Info Provided services to their frequent callers. After Counseling and Info Provided, a significant number of callers are also provided with Victim Advocacy and Safety Planning.

The chart below shows the top 3 services for all of the regions:

Although counseling services are the most recurring services, this chart illustrates how Info provided and Victim Advocacy are not necessarily far behind. CCADV can expect a large portion of their frequent callers to use these services as well, and be aware of this data while distributing their resources.

Uncertainty and Caution

While interpreting our data, it is important to remember that the calls do not necessarily mean a case of domestic violence. Sometimes the calls are made by family members of the survivor, or even the abusers themselves. Some calls are made just to vent or talk about issues, and not because they have been abused. We also must remember that there would be many cases that would have gone unreported. For clarity, we rounded up our percentages and used whole numbers. This affects our data accuracy, but overall does not change any patterns we recognized in this data. One of the most important factors to remember is that the visibility of this pattern does not necessarily mean that it precisely measures the quality or availability of the services. Likewise, the effectiveness of the types of services may vary according to the cases as each case will be unique, and there will be many independent variables such as income, race, family structures…etc. Lastly, the location for some of the callers was not recorded and was grouped under NA. These calls are not included in our data.

Credits

Primarily, we used 2020-2021 masked CCADV data. From those data, we used list callers having more than 10 instances in each region. At the same time, we used records of services provided by CCADV to those frequent callers. Here is the link to CCADV masked data that we used: https://docs.google.com/spreadsheets/d/1dCZftO7ukIlodahZZGwbt5dTAPgenRM6VxO6-Lb_wdQ/edit#gid=1352497557

We used Google sheets to narrow down the major chunks of data we had by creating pivot tables. We also used it in calculating percentages and normalizing our data: https://docs.google.com/spreadsheets/d/1rZ4ndExiHGKMb6H6lFw4aZer-t3OamqDHhWHF755AxU/edit?usp=sharing

Lastly, we used Geojson to download the map of customized regions required for our data set made by Jack Dougherty. We uploaded that map on DataWrapper to make Choropleth maps. We also used a Data wrapper to make charts that pertained to our question.