Perhaps I’ll be the first college student to say this but Morning Afters are the best—

Morning After a new and challenging experience (Of course)–

The quiet time I take the reflect on the events yesterday, before jumping into the current day’s tasks/productivity, is the time that I grow the most as an individual.

As an Industrial Engineer, I have very low tolerance of inefficiency, and at least in my own life, nothing frustrates me more then seeing myself make the same mistake TWICE.

No one’s perfect, so the first time is understandable (as long as I prepared for that the best I could). However if I make the same mistake a second time, that means I didn’t take the “morning after” time to think about what went well and what could have gone better.


Today is the morning after the 2017 Indiana Medicaid Data Challenge.

I heard about this opportunity through an email my IE383, Professor Yih sent to the undergraduate IE students.

In this challenge there was a data analytics and data visualization track. While both tracks were interesting to me, since I have been doing projects on data visualization for sports, I figured:

  1. I could be a more valuable contributor in a team if I brought what I knew/practiced in data visualization to the group
  2. I could able to challenge myself by trying to apply data visualization to the medical industry, which I barely knew anything about

Under Dr. Yih, there were two Ph.D. candidates who I knew because they are my Teaching Assistants for IE 383, and they also were interested in joining the team. From their network, we were also able to find one more Ph.D. IE candidate and a Statistics Ph.D. candidate to make our team.

Our team of 5 entered this contest as “Purdue IE,” and worked tirelessly since Tuesday (after the data release on Monday) to use creative and innovative technology to bring a story to life, through data visualization, that impacts widespread public health. 

The objective statement above is quite significant because there was much debate among our team members on what exactly our deliverable is supposed to look like.

Some questions that we spent discussing were: “How are we to use technology to tell a story?” “What technology?” “What story–do we create an algorithm or is that too analytics heavy?” “Who is our audience and what value does our story add to Medicaid?” “What even is Medicaid?”

Once we made some progress coming to a consensus on what Medicaid even is (a government funded medical insurance for low-income families) we were able to discuss what our deliverable would look like. One of our brilliant team members, Ellen pitched that we create/publish a dashboard that is robust enough to upload any of the data so that there was a way to connect all the random data-sets in the database we were given.

At this point, the idea of “how to use creative technology to visualization” from the objective statement, was complete!


So far, based on our preparations, these is the streamlined process to tackle a data challenge:

  1. Assemble team based on skill sets/credentials needed to thrive in the chosen track (Data visualization)
  2. Make a flowchart seeing how the data in the given database are connected
  3. Define each element of the flowchart/audience to frame how the deliverable will add value to the audience

Moving on to the element of the objective statement that had to do with “telling a story,” our team had a huge chicken and egg debate on whether we should look into each excel sheet (26 in total) first see what elements we find for a story, OR try to make a story line from the flowchart, and then find the supporting data in each excel sheet to complete our story.

Members of either side of the debate had valid points, the data details first–> big picture later people argued that even if we had a great idea if there was no data we could use to support it then it would be useless, so its safer to look at the fine details of the data first.

The big picture first–> data details later argued that we would waste a lot of time digging around with no objective search function, and at least an idea of a story would guide the data digging process.

As a team we finally agreed to look through all the data details first as a team and if we didn’t find anything interesting, we would split up the data among team members and have each team member prepare at least one story line for the next meeting.


Streamlined process continued:

  1. Assemble team based on skill sets/credentials needed to thrive in the chosen track (Data visualization)
  2. Make a flowchart seeing how the data in the given database are connected
  3. Define each element of the flowchart/audience to frame how the deliverable will add value to the audience
  4. Have a high level brainstorm session with all the data sheets, to ensure each member has the same understanding of the data elements, and to try to find a theme/story
  5. If no winning story idea comes from looking at the data details together split the data up among team members and have each individual look closer into their data and present at least one story line, with additional research

Side note: finding a team with smart and hard-working members is really key to make #5 truly valuable

In our next meeting, we brought all our ideas and presented them to the team individually, we narrowed it down to one of the ideas presented and started thinking about how to implement that story on the dashboard.

From there, once the story was identified our time was spent perfecting the visuals in our presentation, making sure every detail like the margins and colors on our dashboard made sense and polishing the verbal delivery of the presentation.

There were only 3 minutes allowed for the presentation so we had to make sure everything we wanted to say was clear and there was a contribution from every element on the visuals to our story.

Finally, it was practice practice practice, practicing the flow of the presentation/the logical structure, and the time (one element that people usually belittle).



Streamlined process continued:

  1. Assemble team based on skill sets/credentials needed to thrive in the chosen track (Data visualization)
  2. Make a flowchart seeing how the data in the given database are connected
  3. Define each element of the flowchart/audience to frame how the deliverable will add value to the audience
  4. Have a high level brainstorm session with all the data sheets, to ensure each member has the same understanding of the data elements, and to try to find a theme/story
  5. If no winning story idea comes from looking at the data details together split the data up among team members and have each individual look closer into their data and present at least one story line, with additional research
  6. Hold a hearing for each individual story idea, narrow down the story to one
  7. Fuse elements of the objective functions (“story” and “creative technologies”) to finalize the product
  8. Remove any distracting visuals from the presentation and review logical structure of the verbal presentation, making sure it emphasizes the event objective
  9. Practice timing the presentation with the speaker/clicker with the correct and final presentation
  10. Present and be proud!

Fast forward to our presentation, I had the pleasure to present our project at the Regenstrief Institure on behalf of my team.

Script/Dashboard link here: 2017 IN Medicaid Data Challenge-Data Visualization (Team: Purdue IE)

We placed first for what would have been the student track and 3rd overall!

Many judges commented on how polished and simple our visualization looked as well as how cohesive the logic of our presentation was.

Most teams had beautiful diagrams and pictures but failed to tell the story in a concise and direct manner, so I’m glad our team made extra efforts to make sure everything we presented and created followed the objective function.

I hope the 10 steps I’ve listed out from will help future teams prep for similar data challenges, and also remind me of our team’s successful approach for my own future opportunities to work with data visualization.


 

Advertisements