CT SafeConnect’s Services Provided and Last Instances

by Michael Kulik and Jack Impronto

Last updated on 01/19/2022

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


When we first were told about CT Safe Connect, we were both incredibly interested in using this data project as a way to help phone representatives in the best way we could. CT SafeConnect is a communication service for the citizens of Connecticut to contact for domestic abuse and other similar issues. Its first recorded data began in 2020 and continues through the present. When given their first two years of caller data, we felt that by determining which actions or advice were given to callers, could be looked at to see which of these actions were most efficient in stopping a problem. We decided to look at only the last numbered instance per person and see what of the seven actions were given in hopes to see a possible correlation or significant change between two or more actions. Although we realize that for many cases there may be more calls in the future yet to come, for these results we are assuming it is the last.

In our research we have found that the CT SafeConnect should be implementing the action of Victim Advocacy when the instance numbers are higher due to the trends we found in our graph of final instances. This would be most important in the Danbury, New London, and Stamford regions as these three have the lowest numbers of Victim Advocacy actions taken per capita.

Victim/Individual Advocacy-”The process of speaking on behalf of an individual or family to ensure their rights and safety” -CT SafeConnect


We started by making a pivot table for each caller ID number so that we could find the max instance number so that we could compare this number with the actions taken on that specific call. We then created a separate table of the instance number and the actions taken. We did this so we could multiply the values of the instance column by the 1 or 0 under each action so that if the third instance was the final and Victim Advocacy help was given, we could mark a three underVictim Advocacy. We were then left with a list of numbers under each action which we used to create a box and whisker plot.

This is our box and whisker graph of the actions taken on the last instance. This type of plot manages to highlight many different details of the spread of data. In the graph below, each of the 7 actions has a rectangle with three horizontal lines. The lowest line represents the 25th percentile, the middle represents the median or 50th percentile, and finally the third represents the 75th percentile. You will also see that above and below the main rectangle there are two lines that extend out. These symbolize the extremes of the data. It is at these points that anything past is deemed an outlier to the data. This aspect of the graph is crucial due to the vast quantity of outliers.

When looking at this graph, there are two main takeaways. First is that there are a significant number of outliers for each of the seven actions. We know right away that these outliers may have a huge effect on any results we may find. Secondly we see that the 25th percentile for Victim Advocacy is higher than the other 6. This finding was what motivated us to move forward looking at solely Victim Advocacy.

This is the Table of times when Victim Advocacy actions were taken for each CT SafeConnect Region regardless of instance number. We decided not to focus on instance numbers to highlight the regions that use this action more than others. We then proceeded to “normalize” these sums for each region. This means that we accounted for the various populations per region and calculated the number of instances per 100,000 people for each region

Region Sum of
Victim Advocacy
Sum of
Instances of Victim
Advocacy per
100,000 People
Ansonia Region 717 122,124 587
Bridgeport Region 2257 322,564 700
Danbury Region 712 231,768 307
Enfield Region 645 106,833 604
Hartford Region 4611 660,846 698
Killingly Region 552 82,6811 668
Mansfield Region 494 84,321 586
Meriden Region 1144 133,419 857
Middletown Region 958 163,053 588
New Britain Region 1741 236,239 737
New Haven Region 3893 449,748 866
New London Region 1284 267,390 467
Stamford Region 1652 379,279 436
Torrington Region 533 106,701 500
Waterbury Region 1837 228,109 805
Grand Total 23660 3,452,951 685

Now we can look at a normalized version of the Victim Advocacy per region making it easier to compare these numbers. We used these numbers and graphed a map of each region so that we can visualize the changes from region to region.

Our map above highlights where Victim Advocacy actions are more likely to be taken per capita. The Meriden, New Haven, and Waterbury regions appear to be most likely to do this. Inversely, the Danbury, Stamford, and New London Regions have the lowest numbers per capita as we previously mentioned.

While it is apparent in the graph that victim advocacy may help ending callers with higher instance numbers, there are many reasons why this might not be true. We started this project in the attempt to see which action on the last instance led to it being the last instance. There however could have been any number of things happening outside this data set that led to a cease in calls for any person. Similarly, just because a call was the last one that was recorded, does not mean it will be the last call ever.

To conclude, we hope that call centers are given advice to utilize the Victim Advocacy actions in hopes it will solve a problem for higher instance callers, specifically in the Danbury, Stamford, and the New London regions.

If you or a loved one is struggling with domestic abuse or are looking for help reach out at any one of these contact methods

Phone Call: 1-888-774-2900

Email: safeconnect@ctcadv.org

Text: 1-888-774-2900

Live Chat: Link available at https://ctsafeconnect.com


Thank you to Maria Guzman and Joanne Vitarelli with Ctsafeconnect for all the Data used.