I-COM Data Science Hackathons 2018

The 2018 Intel Challenge

What marketing strategies influence a brand’s leadership in the Artificial Intelligence (AI) space, and how can data be used to define strategies to engage customers and audiences?

AI as a tool has the potential to transform the future and establish exceptional customer experiences for brands. All over the world, brands are utilizing AI in innovative and interesting ways, striving to become leaders in the AI field. Therefore, a fundamental aim is to develop marketing strategies that advance a brand’s leadership in this field.

The overall objective of the Intel challenge relates to understanding what marketing strategies influence a brand’s leadership in the Artificial Intelligence (AI) space. To better define how a brand can use data to marketing activities we have constructed a quantitative challenge consisting of predicting social conversation volume, and digital engagement for specific topics related to AI. In addition to social engagement volume data, historical cross-platform media investment data will be shared for leading brands in the AI space to aid in the prediction task.

The accuracy of predictions of social and digital engagement generated by each team will be the main quantitative criteria for judging. In addition to the prediction task, each team should create a short presentation where their evaluation, and description of potential relationships between social conversation volume, digital engagement and media investment can be shared. The presentation should highlight the potential and value of accurately predicting consumer engagement with AI topics and brands, and highlight how data can be used to define strategies that can help an AI focused brand better engage customers and audiences.

Challenge Leadership

Prediction task of the Challenge:

Predicting daily social conversation volume, and digital engagement, i.e. page views, for specific topics related to Artificial Intelligence, for a month in 2018, using historical cross-platform media investment data for leading brands in the AI space, and search trends for related topics. Since the prediction task relies on predicting two outcomes with largely different numerical ranges, the accuracy of predictions will be gauged using mean absolute percentage error.

Scoring System
1st Round - 40/60 qualitative/quantitative weight
Final Round - 60/40 qualitative/quantitative weight

Qualitative Scoring for the DSH 2018 Intel Challenge
Each Hackathon team will answer the same questions within each slide deck.
Each question answered contributes to three general categories of scoring as outlined below:
Business Value Storytelling Art & Tech
Main Theme Does this solution solve the client's problem?

Are there additional insights that create added value beyond the predictive problem?
Did the team effectively communicate their methods and ideas? Can the solution scale with bigger data and more problems?

Is the proposed solution elegant, creative and novel?
Scoring 0 - 5 Points:

0 - none present or none sufficiently demonstrated to
5 - high level of business value demonstrated
0 - 2 Points:

0 - not able to communication the value effectively
1 - able to communicate the value effectively
2 - adding more value with excellent communication
0 - 3 Points:

0 - not able to demonstrate unique thinking / is not leveraging technology competitively to
3 - comes out as extraordinarily unique /very surprising use of technology
0 - 10 Points

Challenge Leads