Você está na página 1de 1

Python Data Analysis Exercises

1. Level 1: Get daily returns of Nifty 50 index and calculate cumulative abnormal
return of its individual stocks 7 days before, on the day and 7 days after FY
2018 Budget presentation. Divide the stock CARs into quintiles and back test
return on the spread between the daily returns of bottom quintile and top
quintile stocks over 90 trading days post budget presentation.
2. Level 1: Go to Capexdx database and get the data on stalled projects. They
have data on stalled/stopped/projects under implementation etc. under
different files. There are six or seven files that has different types of data. Some
of the columns are common across all the files. Create two or at most three
files by merging all data by a certain column/columns. Now go to ProwessDx
and download valuation and liquidity ratios of firms whose projects have
stopped yearwise. Merge the two data sets.
3. Level 2 (Requires static web scraping using BeautifulSoup): Go to NCLT
website IBBI orders which contains all court orders of firms dragged to
bankruptcy under IBC 2016. Download all pdf files into four different folders-
admitted, resolved, liquidation and others.
4. Level 3 (Requires interactive web scraping using Selenium): Get a list of BSE
500 firms and their auditors from 2010 to 2018. Auditor data is available on
Prowess. You would have to figure out the list of BSE 500 firms. Ignore index
rebalancing for the time being (although in actual research you would have to
keep that in consideration) and treat all BSE 500 firms at the end of FY 2018
to be the set from 2010-2018. Remove all firms that have more than one
auditor (Most banks, PSUs etc. would get removed from the set and you would
be left with perhaps 150-160 firms). The task is to analyse the returns of firms
around the days whenever an auditor is changed by the company.
A) Flag the auditor changes from the data from Prowess. The database gives
you the year of auditor change. You need the exact date of announcement
of auditor change.
B) Go to BSE website. It has a section on Corporate Announcements. Every
firm listed on BSE has some announcements (AGM, change of directors,
dividend payout, stock splits etc. along with their time stamp, date and
year of when the information was disseminated to exchange). Most auditor
change notifications are inside the AGM announcement document. Using
Selenium package in Python, extract the dates against these AGM pdfs and
put them against the Year of change from the file you cleaned from
Prowess.
C) Get data on daily BSE returns of the stocks in your sample (point B above)
7 days +/- the date of announcement of auditor change. Get the daily Index
returns of BSE 500 for these days (NSE 500 daily returns can be used as
a proxy) and find the CARs (like exercise 1) around the days of auditor
change announcement. (Remember that in exercise 1, the Budget
presentation was a single date. But in this auditor change, dates around
which CARs are to be calculated is different for different firms and is a
moving variable).

Você também pode gostar