Data Analytics Assignment: Text analytics for AirBnB
You have recently got a fantastic job opportunity to be part of AirBnB data analytics
team. As a key member of the team, you’ll be responsible for building the AirBnB
and and
getting, keeping, and developing customers by creating social media contents. You’re using
your expertise in social media. This includes everything from understanding the voice of
customers and creating contents relevant for the company
and to promote in the digital
space and create better customer experience.
At the moment, your team and direct manager do not expect a lot from you.
However, you’re quite ambitious to actively engage and contribute to the team within the
limited budget. One day, you remembered the text analytics lecture from the IMC. As a
starting point, you are going to apply the text analytics technique using Python. The
company has never used such a systemic approach to understanding the voice of
customers. Now, you are about to apply text analytics and theme classification. Further, you
want to upgrade the analysis technique by learning more from the internet.
The success of this project is how much C-level managers can be impressed and
persuaded. Most of all, exciting story-telling using unique text analytics is the key. The C-
level managers have the interest to promote better customer outcomes such as getting,
keeping, and developing relationships with customers, and creating better customer
experiences. The key is that your analysis and proposal must provide clear
ecommendations to help managers achieve these outcomes. Also, excellent written and
visual communication is essential. Finally, be creative and insightful! Enjoy your project!
To do list
1. Apply sentiment analysis and train a classification model to predict polarity of
eviews.
2. Find the most frequent words, show word cloud and visualize the frequent words in
positive and negative reviews separately.
3. Apply sentiment analysis and train a classification model to predict polarity of
eviews.
4. Apply topic modelling and present the list of topics.
5. Discuss the importance of each topic based on the counts and proportions of the
topic out of the entire dataset. In other words, how much does each topic cover the
comments from the dataset?
6. What are the most important issues, opportunities, or insights for the given data?
What are your recommendations to achieve better customer outcomes? List the
ecommendations in order of importance.
7. Provide results in PPT slides and focus on storytelling and visualization. Leaving a
good impression using the PPT slide is important to appeal your idea within a given
short time. Make an interesting and impactful visualization on your slide, also extract
interesting key findings using various analyses.
8. Submit/Upload both PPT slides and Python-files.
Evaluation criteria
1. Impressive, interesting, and informative visualization on PPT slides. (30%)
2. Text analytics capability, depth of analysis & upgraded text analytics technique.
(30%)
3. The quality of recommendations (How interesting/unique/actionable/logical/
persuasive are the recommendations?) (30%)
4. Outstanding overall quality of the report compared to other students. (10%)
Format
• Please use power point (PPT) format (not MS words), because many business
managers prefer to use the PPT to share/communicate ideas.
• Font and size: choose based on your preference.
• Word counts: no limits (please balance between visualization and written
information).
• No animated slides and no video clips.
• PPT page limit: Up to 20 pages including everything. Please note that you’re going to
eport to the C-level management team and they won’t read long pages.
• First page contains the Company’s
and name and your name.
• Second page contains executive summary page, which includes key findings and
ecommendations.