Frankly speaking "so so" tutorial, surprisingly as most other tutorials on DataRobot. It shows "when we click on buttons we see this and this". Excellent! Why to click on them, what the outcome means, what if the process flow, what deployment means. Not structured, not explaining logic. All build on "if you click here you see the graphs and text". How to actually get the answer from initial question "how to get a list of people with diabethes"? When I click "deploy" - it shows different interface, and asks to fill in feature, date format, application ID, and a few toggles. The prediction ghraph is not clear at all as well as its explanation. If I would to buy something for $100k per year - i would be highly motivated to see something more clear, structured and consise. Sorry, the efforts are visible, but probably only for people who already understand everything how it works. but not good enough for a newbies.
Thanks Prof. Rayan! It is really a nice tool. However, I have some concerns: 1- Does it work with data that comes from various data sources such as SQL database and API, or only it works with Excel files? 2- Does the model retrain itself automatically when it is fed with new data entries? 3- Can all of the mentioned steps be automated as soon as the data source updated? 4- Does it allow data wrangling and preprocessing? 5- Can we use different data sources and join them inside DataRobot? For example, let's say we have demographic data about patients from Excel sheet, and we also have the diabetes data in a way you showed us in this video, so can we join them and create a new dataset before we insert it into training? I mean something similar to JOIN in SQL, or Merge in Python, or the popular Vlookup function in Excel? From my point of view, I believe the tool is an amazing solution if it ensures the updating, automation, and collaboration between different data sources.
What happens if I have to do some data transformation in order to improve the Model? For example, some time I have to add some interaction between two features or combination between them, some like Y = X1 + X2 + 2*X1 + (sqrt(X2))? How the Datarobot will work on this case that I have to do some enhancements on the features or add some new features (from the features that already exists)?
I read that DataRobot costs 30k a year, I can see that the value is amazing but is the price in any way related to the business size because 30k is still a whole lot for a small business with 3 people.
Frankly speaking "so so" tutorial, surprisingly as most other tutorials on DataRobot. It shows "when we click on buttons we see this and this". Excellent! Why to click on them, what the outcome means, what if the process flow, what deployment means. Not structured, not explaining logic. All build on "if you click here you see the graphs and text". How to actually get the answer from initial question "how to get a list of people with diabethes"? When I click "deploy" - it shows different interface, and asks to fill in feature, date format, application ID, and a few toggles. The prediction ghraph is not clear at all as well as its explanation. If I would to buy something for $100k per year - i would be highly motivated to see something more clear, structured and consise. Sorry, the efforts are visible, but probably only for people who already understand everything how it works. but not good enough for a newbies.
ممتاز
Thanks Prof. Rayan! It is really a nice tool. However, I have some concerns:
1- Does it work with data that comes from various data sources such as SQL database and API, or only it works with Excel files?
2- Does the model retrain itself automatically when it is fed with new data entries?
3- Can all of the mentioned steps be automated as soon as the data source updated?
4- Does it allow data wrangling and preprocessing?
5- Can we use different data sources and join them inside DataRobot? For example, let's say we have demographic data about patients from Excel sheet, and we also have the diabetes data in a way you showed us in this video, so can we join them and create a new dataset before we insert it into training? I mean something similar to JOIN in SQL, or Merge in Python, or the popular Vlookup function in Excel?
From my point of view, I believe the tool is an amazing solution if it ensures the updating, automation, and collaboration between different data sources.
What happens if I have to do some data transformation in order to improve the Model? For example, some time I have to add some interaction between two features or combination between them, some like Y = X1 + X2 + 2*X1 + (sqrt(X2))?
How the Datarobot will work on this case that I have to do some enhancements on the features or add some new features (from the features that already exists)?
I read that DataRobot costs 30k a year, I can see that the value is amazing but is the price in any way related to the business size because 30k is still a whole lot for a small business with 3 people.
How much does it cost to use this DataRobot?
DataRobot tool is easy to learn
3:23