I am currently writing a diploma thesis on predictive analytics in intelligent buildings at Prague University of Economics and Business. I would like to use the Snowflake tool and your DataRobot modeling tool. However, in the trial I encountered several limitations that do not allow me to complete this thesis.
Therefore, I would ask you if it is possible to use your trial to the fullest? This is mainly about time series modeling, which is not allowed in the trial version. I have an extended trial version until December 15, when the deadline for submitting the work. For example, I have data on people walking into the building and I want to find out how many come in a week. Is there any alternative functionality in trial version?
It would have two benefits for you. The first one is to raise awareness of this platform in a public publication. The second one is the fact that if these models and DataRobot platform will be great for the company where I am implementing this solution, then Sharry company - https://www.sharry.tech/ will begin to use your services and you will get a new customer.
Hi @EtheLe!
Nice to meet you. I'll reach out to you via email to answer all your questions.
Iryna
@EtheLe, your right the Trial version doesnt have AutoTS and out-of-time validation isn't a forecasting method. To get an enterprise version of the software you could contact DataRobot via the commercial website. If you want a demo of AutoTS or talk about your project feel free to email me at watt@bulien.com.
@IraWatt Thank you for your reply. But I think i haven't access to this time series forecast modeling. For example i didn't see anywhere this setting:
I think in trial version I can't predict future. I have only these settings:
If i create this model with Out-of-Time Validation i haven't in result models future prediction. Can I create forecasting in trial version by this version of Data Robot? What model can return me future numbers of employees arriving?
Hi @EtheLe, thought I'd link this article from the DataRobot community, very similar to your use case.
Thank you for your response. I tried this method directly according to the instructions in the documentation, but I did not find a forecast of the future there. Only prediction results on historical data. I need to get a forecast for example for the next 14 days:
date_arrived | number of employees
2021-05-01 | 20
2021-05-02 | 52
2021-05-03 | 30
...
2021-10-11 | 120
From this historical data, I need to make a model that in the final prediction will have a future forecast of employee arrivals:
date_arrived | number of employees
2021-10-12 | 43
2021-10-13 | 35
I did not find a future in individual models, only past data. Can I do this only with Automated machine learning and Out-of-Time Validation? Without access to Automated time series forecasting.
Hi @EtheLe - You can accomplish many time series use cases using OTV (out of time validation). The is explained in the public documentation: :Date/time partitioning . Have a look and let us know if you have additional questions.
And, wishing you the best on your diploma thesis. The DataRobot public documentation has a lot of great content that may help in your journey.
linda