These are chat archives for FreeCodeCamp/DataScience

24th
May 2017
Himujjal Upadhyaya
@Himujjal
May 24 2017 08:35
@NehaSehrawat https://github.com/Himujjal/Awesome-Hacking - Check this repository. This may help you
satenndrra
@satenndrra
May 24 2017 10:31
@PunkDado I thought that the 2017 data on github repo was already cleaned and in CSV format. Are you doing some additional processing?
Neha Sehrawat
@NehaSehrawat
May 24 2017 10:35
@Himujjal thanks I'm on it!
CamperBot
@camperbot
May 24 2017 10:35
nehasehrawat sends brownie points to @himujjal :sparkles: :thumbsup: :sparkles:
:cookie: 321 | @himujjal |http://www.freecodecamp.com/himujjal
ridaamine
@ridaamine
May 24 2017 11:43
@evaristoc thanks for your reply ans suggestion i'll look into the advice you gave me
i'm now trying to fit an SARIMA model i tested with the Dickey fuller test if the series is stationary and it is and plotted the ACF/PACF to determine the SARIMA parametres but i'm kinda stuck here didn't understand how they determine them with this
here's the plot
CamperBot
@camperbot
May 24 2017 11:43
ridaamine sends brownie points to @evaristoc :sparkles: :thumbsup: :sparkles:
:cookie: 352 | @evaristoc |http://www.freecodecamp.com/evaristoc
ridaamine
@ridaamine
May 24 2017 11:44
ACF & PACF.png
evaristoc
@evaristoc
May 24 2017 12:18

@ridaamine It seems to me that both charts are showing stationary periodicity with a lag of 0.5, whatever your lag metric is.

Assuming everything right and I am correct, you might need a function in your model that reflects that periodicity.

Disclaimer : Be aware that I have applied time series to solve a couple of problems before but I am not an expert.

Additionally...
A possible reference? There are many online, this was the first I found from a quick search. With R scripts.
https://www.stat.tamu.edu/~suhasini/teaching673/time_series.pdf

ridaamine
@ridaamine
May 24 2017 14:23
@evaristoc sorry to bother you more but reading the ACF/PACF part i still don't get how they can from these two graphs determine the parameters in the ARIMA model
Leonardo Raduy Lemos
@PunkDado
May 24 2017 15:02
@satenndrra , data is cleaned. I'm doing analysis. The files I'm generating are sub-datasets of the cleaned data.