Autocorrelation plot python interpretation for time series forecasting?

Question

I'm willing to predict a variable x, my data is private data so I cannot share but we have data time as index and one only column with the variable in %:

I didn't know what machine learning method to use so I started with time series forecasting. I'm using this guide and doing the following:
https://towardsdatascience.com/your-comprehensive-guide-to-the-basics-of-time-series-modeling-f673398b5df3

Now my problem is that I'm not understanding a graph, the autocorrelation_plot . I used the pandas autocorrelation_plot(d) :

I'm not understanding the output of this graph after hours looking it up on several sites. Could someone help me?

Also, would you suggest a different method for prediction for this kind of data?

Thank you !!

Answer 1

The autocorrelation plot represent the correlation between each term and itself with the lag in the x axis.

Looking at your plot you can see that begins at 1, because this represents the autocorrelation between each term and itself.

Going further you can see that the autocorrelation increasing the lag decrease, until being near zero.

Passed an year of lag (assuming that the lag is in days) we can see that this autocorrelation decrease more until going again to zero.