Yelp from business perspective
Popularity of Yelp over time
The popularity of Yelp is driven by two entities: Users and Businesses. It can be measured by the number of reviews as it involves both the entities. I did a time series analysis to find the trend and seasonality of the Yelp activity in Edinburgh. I used review count for each month to plot the following graph
Exceptional peak in the number of reviews is observed between 2010 and 2011, which makes it difficult to infer the overall trend. To overcome this I applied log transformation, which penalises the higher values more than the smaller values and it shown as Original time series in the following figure.
This transformed series in decomposed to find Trend and seasonality. Since trend and seasonality are already giving us a good picture of the popularity so there is no analysis is done on residue.
Analysis And Inference
Activity in Yelp was highest between 2010 and 2011. To investigate the probable reason behind, I checked statistics of Yelp and Edinburgh. In 2010, Yelp was introduced in 4 European countries: France, Germany, Austria and Netherlands. Also, “Overseas worker registered in Edinburgh 2003 -2013”, shows an increasing number of European migrant after 2009. It might be that Yelp popularity was influenced further by the influx of European migrants who were already active Yelp users. It might be interesting to find the real cause of the spike, but the data made available by Yelp was not sufficient.
The overall trend is increasing from 2008 to 2011. After a dip, it has become constant. Above figure shows the seasonal effect of reviews count. To study it in detail, I took the snapshot of the same seasonality graph from the year 2011.
It can be observed that December, February and September are the calmer months, whereas July, January, April and October are activity months.
Next Post I will Compare activities between different business Categories in Yelp.