Eat, Eat, Eat

Visiting Openrice website and doing assignment really makes me feel hungry，but I still have to think.You can visit my website here .

My working experience

I focus on most favorite restaurants in Hong Kong Island for data extraction and I analysis the existing restaurant market performance and try to find out what types of restaurants and what locations that may enables the restaurant to have the opportunity to earn money. I choose the data in picture as data columns. The extracted data is relatively neat, I just need to remove the brackets around the number of food reviews and combine the address as well as replace the‘粤菜（广东）’ with ‘粤菜’,etc.

After I imported the database in Jupyter notebook, I first count types with a column, then display data of one categories using group by, and then analyze the correlation of some selective data, the details like calculating their average positive feedback rate and compare it with the overall positive feedback rate.

An interesting discovery while visiting the food review pages though not proved by data.

I am interested in users who produce content posting habits and behavioral motivation However, I search the Openrice’s robot text, in which I found that user information is protected; based on my observation without data support, high-level users who write a lot of food reviews will be more attentive to put a lot of photos and attach text, while users who write a small number of food reviews are more likely to give a bad rate to the restaurant, and they come up to share because of dissatisfaction.

Questions remained

I found that when the rating criteria such as ‘positive’ and ‘negative’ and ‘ok’ are expressed using images with no text description, so the data cannot be grabbed, how to get this data?