The objective of PurpleTruth is to provide a news application that includes an intuitive yetinsightful measure of political bias for readers as they are consuming their current news sources.Our goal is to create a machine learning algorithm that can produce a bias indicator, which wouldbe red (conservative), blue (liberal) or purple (neutral), based on natural language processing andcrowdsourced feedback for any article online that’s from a verified news source. Customers wil be able to see this bias indicator on all versions of the PurpleTruth application, and on any newsaggregator or website that integrates with PurpleTruth via APIs.
Please find a prototype walkthrough video here: https://youtu.be/Paxwzn2Lgp0
The components of the MVP include:-
- A free news aggregation mobile application (“PurpleTruth”) with a “Bias Tracker” biasindicator feature. For our MVP, we will focus on breaking news and U.S. political news(while excluding opinion pieces). Our application will present stories from different newssources with a bias checker on each article to enable the readers to read articles fromdifferent bias vantage-points.
- A “Bias Check” API that will be made publicly available at a cost. This API will allownews services and aggritators to present a red, blue or purple bias indicator and a “reportbias” prompt that will be displayed next to article names on news applications and sitesthat integrate with our service.
- A web page hosted by PurpleTruth that will utilize the “Bias Check” API for ad hocrequests by users to verify the biases of a news article upon submission.
- A Machine Learning Algorithm based on Natural Language processing will be the coreengine driving our Bias Check API and the Bias Tracker on the PurpleTruth application.
- In-house team leading newsarticlesand news source verification, as well as analysingcustomer feedback, to ensure quality of our Bias Tracker.
- Full transparency into all internal (human based) processes, independently verifiedthrough random audits conducted by an objective third party.
Steps taken:
-
Ideation and validation of usecase using user interviews and survey. The results can be found in the report above.
-
Wireframing using balsamiq. Please find the user test versions of the wireframe below: Application: https://balsamiq.cloud/sjmgd6i/p7bljgx/r1C0F?f=N4IgUiBcAMA0IDkpxAYWfAMhkAhHAslAIzwBaOA0lANoC6AvkA%3D%3D Webiste: https://balsamiq.cloud/sjmgd6i/p7bljgx/r8B3C?f=N4IgUiBcAMA0IDkpxAYWfAMhkAhHAsjgFo4DSUA2gLoC%2BQA%3D
-
Prototyping using Adobe XD. Application: https://xd.adobe.com/view/2a4d8c1c-ff17-49b0-b214-91d6645858c9-7f1e/?fullscreen&hints=off Website: https://xd.adobe.com/view/ea2e2f09-df6a-4409-bcb9-ea6fa7d1429b-bf7d/?fullscreen&hints=off
-
User tests on the prototype and interative improvements
-
Pitching the idea.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent