Our Deliverables

We had an amazing time presenting our project at the Georgia Assembly last night. With only one week left in the program, we will be focused on making sure our deliverables are ready to hand off to Don Zubler, the Operations Director at the 211 Center. Our deliverables are illustrated below.

DeliverablesThe first three deliverables are already completed; we simply need to put some finishing touches on these. The final deliverable, the report, will encompass all of  the research and analysis we have done for the past nine weeks. Using this analysis, we will present a series of suggestions on potential improvements and optimizations our partners at 211 can use to improve the services they provide.

Online Interactive Application

Over the course of this project we have implemented several programs in R for analyzing different aspects of the data available to us. All of the analyses done so far using these programs are based on the data collected previously. In order to apply these analyses on future data, one needs to install R software on its computer and rerun the programs with new data. This might create some difficulties for the user. Therefore, we decided to build interactive applications which can perform certain predesigned analyses on the data set given to them. Data can be uploaded to the designated application and can be analyzed after processing the algorithm. Moreover, these interactive applications can be hosted and accessed through the web which can be accessed on any device without having to install any software. They can also be upgraded to read the data directly form the user database and then perform the intended analyses, if permitted by the database owner. An example of such application is shown below.

 

Blog

Analysis of Calls per Zip Code

Atlanta’s 211 service reaches out to several counties and zip codes in the state of Georgia. In order to analyze its outreach in Georgia zip codes, we analyzed the call volume for January 2015 – May 2015 for each zip code. The call volume for the top 15 zip codes is graphed below.

Zip Code

 

We plan on using D3 to design an interactive visualization of the number of types of call per zip code. For example, someone will be able to see how many people call for bill assistance in the 30349 zip code.

 

By analyzing call volume per zip code, we can also analyze why certain zip codes have fewer calls than others. The reason could be as simple as population differences. However, it could also be a lack of awareness of 211. Therefore, we will use this analysis to recommend potential advertisement efforts.

211 Data Visualization

Seeing that the mid-project review is next week, we have really been focusing on creating visualizations of the data we have collected so far.  These visualizations will not only help us present the findings of our project but also help us propose changes to the call menu based on our analysis.  Analyzing the data will also help us create a predictive model for the number of calls for a given day of the week, month, and year.

 

Calls per Year

This is the number of calls per year from 2011-2014.  The decrease in the number of calls is due to the introduction of email, text, and chat services.

 

Calls per month

This is the number of calls per month from 2011-2014.  Peak months include the summer, as people look for assistance with childcare and summer camps, and October, as people seek assistance with purchasing gifts for the holidays.  February has the fewest calls, as many people receive income tax refunds in January and therefore need less assistance the next month.

 

Calls per day

This is the number of calls per day for each day of the week.  Most calls occur on Monday and decrease as the week goes on.  This is because people prefer to seek resources earlier in the week so they have more time in the week to explore the options the agents direct them to.

 

Call dropped time

This plot illustrates the time callers spend in the menu before hanging up.  Most callers abandon the call between 5 and 10 seconds, during the first spiel of the call menu.  The percentage of abandoned calls hits a local maximum around one minute, which is during the second spiel of the call menu.

 

Weather conditions

This is a plot of the number of calls received in 2013-2014 based on weather conditions.  Weather conditions such as snow, fog, and thunderstorms increase the number of calls received.

 

 

Answering the Call of Duty

We have just wrapped up our second week working with the United Way of Metro Atlanta’s 211 Call Center (for more information on the 211 call center, check out last week’s blog post!).  This past week has been very productive and given us more insight into the problem we’re trying to solve this summer.

 

After analyzing some of the sample data we gathered last week, we decided to collect more data about abandoned calls.  Looking at the data from the past few months, we noticed that there are a few numbers that have called hundreds of times a month.  We will notify the 211 director of this issue.

 

Our main goal for the summer is to analyze the data to make a menu that benefits both the callers and the agents.  The current menu is pictured below.  Currently, the menu is long and repetitive with some inaccurate prompts.  Some of the improvements we hope to make on the caller’s end are condensing the information, taking out the repetitive sections, and allowing repeat callers to skip information they already know.

Call tree

On Wednesday, we had the opportunity to experience the other end of calls by listening to agents handle calls.  This was very beneficial, as it allowed us to see what agents do for each call and what they would like to be improved.  The biggest potential improvement involves data entry.  All of the information gathered is manually entered by the agent while they are on the phone.  Agents have to rush to input the caller’s age, zip code, insurance status, employment information, and more while trying to find the best organizations to handle the callers’ needs. A way we hope to improve this is having callers input numerical data (phone number, age, zip code, etc.) and yes/no questions (veteran status, insurance status, etc.) before the call is connected to the agent.