search

rpasquini / twitter_and_displacement

Predictive analysis of gentrification and displacement on the basis of geo-tagged tweets based metrics.
2 stars 5 forks source link

Econometrics by typologies #2

Open rpasquini opened 4 years ago

rpasquini commented 4 years ago

In order to test the relationship between typologies and twitter growth we will implement an econometric analysis:

image

where dtypo_j refers are dummies for typology j and t denotes a linear time trend

rpasquini commented 4 years ago

Here is an example of the regression analysis for Buenos Aires [here].(https://github.com/rpasquini/twitter_and_displacement/blob/master/notebooks/Econometrics%20by%20typologies%20Buenos%20Aires.ipynb ). I will add the process of data preparation (aggregation of counts to the tract level) to the main step by step guide. This data should be stored on the server. I will further propose we standarize the names of the main variables so we can specify the same regressions across all cities

rpasquini commented 4 years ago

@Emman-Lopez-Oso please let me know in which cities we have the hex counts ready (and the typologies data available) in order to advance in this analysis.

Emman-Lopez-Oso commented 4 years ago

@rpasquini Sydney, Australia and Hong Kong, China

We have Bogota's hex counts pending, and San Francisco and New York were unable to calculate the hex counts because of issues in the data.