Yelp / Weather Hackathon

Discussion Questions

Data

1. Review count, and negative/positive reviews may be affected by weather for businesses. There may be more positive reviews when the weather is really nice out, or vice versa.
2. No
3. You could possibly predict the weather based on the reviews found in the Yelp dataset.

Method

1. Temperature, time, review count
2. You could find keywords that may correlate with the trends in the weather dataset

Bias

1. Yelp dataset only has information for three cities.
2. You'd only be able to get positive reviews
3. We would state that the dataset sample is not representative of its population.

Objective 1

Rainy 3/10,3/11,4/23,4/24,4/25

Sunny 2/8, 2/10, 2/15, 2/21, 2/23, 2/24

Objective 2

1. The businesses could all be put on a map, where size of the circle represents number of the reviews and color of the circle is average review. A new map could be created for each day, and the visualization would update the map for each day, sort of like a time lapse. To represent the weather, there could be storm clouds at the top of the visualization representing rain, or sun rays representing sunshine, etc.

2. It's not enough to simply show the businesses with number of reviews for each of the selected days because we're trying to find correlations with weather, so the weather also has to be tied into the visualization.

Objective 3

[Tell your story! Link your graphics!]

Objective 4

[Do it again!]