This project is based on sentiment analysis on twitter + algorithm based on the moving average algorithm. The first step of this algorithm is to create two curves :
- Short moving average curve (2 periods in this case)
- Long moving average curve (20 periods in this case)
The two moving average curves are based on the closing prices curve on the given period (1 day or 1 hour). When the short moving average curve is crossing the long moving average curve from below then it is time to buy stocks. When the short moving average curve is crossing the long moving average curve from above then it is time to sell stocks.
In our case, we find out that the moving average combination 2/20 seems to be optimized for cryptocurrency.
For example, here is a plotting of the closing prices for the Bitcoin. There is also the MMC curve (2 days moving average) and the MML (20 days moving average).
It this plotting you can see that the moving average algorithm advised to sell on 2021-04-18 because the short moving average curve (MMC) is crossing the long moving average curve (MML) from above. This algorithm has been a bit enhanced in this application because trend volume, the derivative of the long-term moving average and the spread between the short and long term moving average curves are also considered.
In this application, there is 2 cases:
- daily alerts
- daily trend recommendation
The daily alerts program is running our moving average algorithm every hour with a period of time of 1 hour. If one cryptocurrency has to be sold or has to be bought then an alerts is automatically sent to the correct discord channel.
The daily trend recommendation program is running our moving average algorithm with a period of time of 1 day every morning (8:30) and every evening (17:30). Recommendation is sent for every tracked cryptocurrency in the right Discord channel.
When the short term moving average (MMC) is higher than the long-term moving average (MML).
| Situation | Recommendation |
|---|---|
| If the long-term moving average (MML) is increasing and if the spread is increasing. | In a up-trend. You can buy ! (it might be too late) |
| Else | Wait for sell.. |
When the short term moving average (MMC) is lower than the long-term moving average (MML).
| Situation | Recommendation |
|---|---|
| If the long-term moving average (MML) is increasing and if the spread is decreasing. | In a up-trend. You can buy ! (it might be too early) |
| Else | Wait for buy.. |
When the short term moving average (MMC) is crossing the long term moving average (MML) from below.
| Situation | Recommendation |
|---|---|
| If the long-term moving average (MML) is increasing and if the volume trend is higher than 50% | Buy Today!! (or Buy Now!!) |
| If the long-term moving average (MML) is increasing and if the volume trend is lower than 50% | You can buy but it's not a volume trend. |
| Else | Wait. |
When the short term moving average (MMC) is crossing the long term moving average (MML) from above.
| Situation | Recommendation |
|---|---|
| Case 4 fulfilled | Sell Today!! (or Sell Now!!) |
This part is requesting data to the twitter trend api project. It is an algorithm based on NLTK which is really basic but is will be enhanced in the near future.