- Wang Jin & Zhongjie He
The aim of this project is to pick the public companies(have listed on American stock exchanges) from the original 222,450 companies(some are repeated). The file work1.csv is the work file and the work2 is the public companies(more than 6,000) in United States' share exchanges. Mainly we apply fuzzy wuzzy algorithm and picked the required ones. https://github.com/seatgeek/fuzzywuzzy (developped by SeatGeet)
**There are two main difficulties. **
- The first names in the file with unchecked companies are not very formal, say 'Apple company ' instead of 'Apple Inc'. We designed the fuzzy wuzzy value by making experiments manually in advance.
- The second is massive data leads to much running time (it is estimated more than 12 hours for the first version programe). We reduced it to less than five minutes by devide the matching procedure into two steps after carefully observed the data structure.