When I travel to my home country I always unsure what the best date is to exchange my money to get the best possible rate. It is even more difficult in case I need to exchange a larger amount. Other people might be in the same situation when they need to exchange money, so in order to make the decision easier I have created this Flask based API using machine learning, to find the best date in the near future, and created a web interface to access the API functionality, which can be checked here.
- EUR -> BGN, GBP, HUF,PLN
- GBP -> HUF, PLN
- USD -> GBP, HUF, PLN
- GET /currencies
- Returns the available currencies as a JSON
- Endpoint: http://laszloszoboszlai.me:5000/currencies
- POST /tocurrencies
- Returns the available currencies to translate from a specific curency
- request format : { "currency" : CURRENCY }
- Endpoint: http://laszloszoboszlai.me:5000/tocurrencies
-
POST /forecast
- expected JSON format :
- {
- "currencyfrom" : "CURRENCY",
- "currencyto" : "CURRENCY",
- "days": X
- }
-
Response:
- {'lasknownrate' : last known exchange rate (date)
- 'accuracy' : accuracy of the prediction (float)
- 'forecasts' : forecasted exchange rates (array of dictionaries of date:rate pairs for easy plotting)
- 'tosell' : the predicted day for selling the currency (date)
- 'tobuy' : to predicted day for buying the currency (date)
- }
- My plan is to create another API for sentiment analysis of different financial websites to improve the prediction accuracy from the gained knowledge.
- Create a mobile application using the API.
- Move to microservice architecture.