Stock market.
Investors and Traders in stock market.
By analysing the given data we help investors and traders to make buying and selling decisions in a way to gain an edge in the market.
Currently, to make any investment decision we are relying on stock brokers who provide us the current trends happening in the market which include their opinions. Based on their inputs we invest/buy or sell shares. With our approach we can minimise the role of stock brokers. This helps the users to know the trends in the market and also make unbiased decisions.
● Data Set: ○ The data set consists of two csv files - “instruments.csv” and “log.csv”. ○ The “instruments.csv” file contains attributes which describe about various stocks. The “log.csv” file contains transaction data based on the depth and timestamp of the stocks. ○ The “log.csv” file contains transactions that were recorded for a period of three days. It also has some depth parameters which show the supply and demand for stock at various prices and can be an indicator of market sentiment.
● Initial Assessment Tools and Techniques: ○ R ○ Tableau
● Exploratory Analysis : Find Support level and resistance level of a stock. Support level is a price level at which the stock might find support and below which it may not fall. In contrast, a resistance level is a price at which the stock might find pressure and above which it may not rise.
● Descriptive Analysis : Summary of a particular stock based on the selling and buying quantity of that stock. This can be shown in the form of bar charts, box plots, pivot tables, etc.
● Classification : Based on quantity of a particular stock sold compared to quantity of that stock bought in a particular time period, we classify it as a buying trend or selling trend.
● Clustering : Clustering can be done on various attributes like strike rate, average price, sell price, etc. This can help us in forming clusters which tell us about the profits or budget. The dataset has some depth parameters too. Clustering can be done on them to obtain some results.
● Association : If the values of the raw inputs required for a particular company change (increase or decrease) then it will affect the share price of that company. For example, if rubber becomes cheaper, then the manufacturing price of tyres will fall leading to increase in the prices of shares of a tyre producing company. Association can be possible after we obtain some clusters as described before.
● We assumed that this is the real world data. The data gives Real-world data for half of the Nifty 50 stocks for last week of May and first 2 Weeks of July.
● All the data points are not tick-by-tick update. Rather it is mostly an update after 600 ms, provided a trade happened.
In the collected dataset, information about the stock like to which sector (consumer-goods, IT sector, financial,banking) it belongs is not given. So, analytics based on sector may not be possible.
This type of data analytics comes under Technical analysis which assumes that, market value of a stock depends solely on the supply-demand which is reflected in the past trading patterns. So, our inference model is only capable of predicting the support and resistance values based on these trends. But, in real-world, investments are subject to market risks. They depend on various other factors like sentiments, reputation, politics etc.
https://www.kaggle.com/deeiip/1m-real-time-stock-market-data-nse/data