Data Visualization

 

Literature Insights / Data Visualization Exploration

Distribution of evidence over time

Goal of Visualization

Help the user understand the breadth of literature available on a specific topic at a glance so that they can see trends in the evidence over time and dive deeper into the papers with high expert scores or new findings reported.

Click images to view larger 🔎

Dot Histogram A

Pros: Easy to see trend of data over time, No y-axis means no clustered data

Cons: No additional data about the publications

Dot Histogram B

Pros: Y-axis provides additional information about quality

Cons: Not as easy to see trend at a glance, Clustered data

 

User Feedback

We heard from oncologists that understanding the quality of evidence is an important piece of data when evaluating which papers to read. This meant that Dot Histogram B’s ability to provide that additional layer of data on the Y-axis made it more valuable.

 

Relevant treatments by volume and quality

Goal of Visualization

Help the user understand which drugs have been studied for their target cohort and the quality of the published articles where these drugs were studied to inform their evaluation of different treatment options.

Click images to view larger 🔎

Bubble Chart

PROS: Quick summary of relevant publications

CONS: Difficult to display treatment names, Question value of combining quality of papers into an average, Relies on color to distinguish between the quality categories

Stacked Bar Graph

PROS: EASIER TO COMPARE QUANTITY OF PUBLICATIONS, LABEL CATEGORIES OF QUALITY, EASIER TO DISPLAY TREATMENT NAMES

CONS: RELIES ON COLOR TO DISTINGUISH BETWEEN QUALITY CATEGORIES

 

User Feedback

We heard from oncologists that the additional layer of quality average or quality breakdown per treatment was not something that would help them make decisions. Instead, they expressed interest in a chart that visualized how the treatments compared (see next section).

 

Relevant treatment outcome comparison

Goal of Visualization

Help the user understand which drugs have been studied for their target cohort and how they compared against each other so that they can confirm or build their treatment hypotheses using the available data.

Click images to view larger 🔎

Arc Diagram

PROS: EASY TO SCAN FOR TREATMENT NAMES

CONS: LIST VIEW MAY IMPLY HIERARCHY, TAKES UP A LOT OF VERTICAL SCREEN REAL ESTATE

Chord Diagram

PROS: ORGANIZE TREATMENTS BY DRUG CLASS, A LOT OF DATA IN SMALL SPACE, ADDITIONAL LAYER OF DATA (# OF PATIENTS STUDIED)

CONS: NOT EASY TO SCAN FOR TREATMENTS

Network Diagram w/ Arcs

PROS: EASY TO VIEW CLUSTERS OF TREATMENTS WHICH HAVE BEEN COMPARED OR NOT COMPARED, LABEL BY DRUG CLASS

CONS: less common DATA VISUALIZATION

User Feedback

We heard from oncologists that comparing outcomes was a more valuable way to summarize relevant treatments. By providing a snapshot of not only what drugs had been compared, but also which drugs had yet to be compared, we were providing a quick insight into the literature’s sentiment. One piece of feedback that applied to all explorations was that it might be valuable to add another layer of data to the outcome “arcs” – what outcome was being studied? Overall survival, progression-free survival, toxicity, response?

Overall, user feedback indicated the network diagram with arcs seemed to be heading in the best direction. Because the network-style visualization forms clusters it is a bit easier to see which drugs have been compared/not compared.

Unfortunately, this visualization was not able to be included in the product beta due to implementation complexity and the accelerated project timeline leading up to the ASCO conference. Moving forward, the development of this visualization will be re-sized to be introduced into the Literature Insights experience.

 
 

return to Literature Insights to see the project solution