Rapidminer studio to integrate with IOT development bord

arbitary32arbitary32 Member Posts: 3 Contributor I
edited November 2019 in Help

Rapid Minor studion is new for me, Isi t possible to install this siftware in raspberry pi?

Basically I need to perform some task with Wireless Sensor which sends the reading to create the tabular form using rapid minor studio,

But as I am beginner to this area, So need suggestion how to integrate the readings of sensors to stored in tabular form and store it on cloud database using Rapid Minor 

 

Your Leads will be much appreciated 

 

Note: I am using Losant API to store the the readings in database 

Answers

  • jczogallajczogalla Employee, Member Posts: 144 RM Engineering

    Hi @arbitary32,

     

    in general it should be possible to run RapidMiner on a Raspberry Pi, since it runs on Java. There is a platform independent version you can download from the website.

    However, I think in your case it would be easier to just use your Pi to store the sensor values in some database and then access that database from RapidMiner, running it on a normal computer. RapidMiner can handle a lot of different JDBC connections, you can read how to use  MySQL or MariaDB here.

     

    Cheers

    Jan

  • sgenzersgenzer Administrator, Moderator, Employee, RapidMiner Certified Analyst, Community Manager, Member, University Professor, PM Moderator Posts: 2,959 Community Manager

    tagging my RPi guru @Thomas_Ott

     

    Scott

     

     

  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    @arbitary32 You can run RapidMiner Studio on a Raspberry Pi but it will be woefully slow. The start up time using the version 2 of the Pi (512 MB ram) was like 5 minutes. I haven't tested it on the version 3 boards, so maybe you'll get better performance BUT I suspect it will be slow.

     

    If you want to use the Pi as an IoT board, use it collect the data. Pipe it a database and then use RapidMiner Studio on a more powerful machine. 

     

    Of course, you can get an OEM licenses and embed RM java classes directly, that might be feasible but I haven't tested it. 

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,503 RM Data Scientist

    Hi,

     

    a better way to integrate are the new RapidMiner Scoring Agents. They should work on a Pi.

     

    Best,

    Martin

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member, University Professor Posts: 568 Unicorn

    Hi @arbitary32,

     

    To work with IoT, you do not need to install RapidMiner Studio on your Raspberry Pi, it's RapidMiner Server. Let's put it this way: you can build your processes on RapidMiner Studio and deploy these on RapidMiner Server. Either way it isn't a great idea to install a RapidMiner product on Raspberry Pi. I did it once on a NUC and totally regretted that decision later.

     

    Now, I have some experience working with Losant Klin and its RESTful API's. Given that, I can provide you a scenario that IMO might be better: install RM Studio on your computer and RM Server somewhere else (if you have enough memory to run both on your computer, that's great, but my MacBook Air can't cope with it). You can build processes on RapidMiner Studio, deploy these on the Server (it's a copy-paste process, don't be afraid!) and expose these from the Server as API endpoints (it's a simple configuration done in the server interface) that you can call with any HTTP-capable client from your Raspberry Pi boards (curl is a good start!).

     

    The interesting part of this setup comes a bit later, though: if you want to simplify your life, you can let RapidMiner handle the information sent to the Losant API, by using the RapidMiner Python extension from the Marketplace and the Anaconda Python Distribution. You just need to install the library found on this repository and write your code in the Python Programming Language as part of a process from RapidMiner Studio/Server. That way, the amount of code you need to write specifically for your Raspberry Pi is significantly less and the possibilities are endless!

     

    All the best,

     

  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    @rfuentealba that's a pretty good idea to have the Pi hit the Server exposed REST API. I think it's what @mschmitz was alluding too with the real time scoring product of the Server. 

  • arbitary32arbitary32 Member Posts: 3 Contributor I

    @rfuentealba is it possible to create the the dashboard in line graph and display the dashboard thorugh smartphone whenever  smartphone is connected with internet and also share the notificatio alert whenever temprature drops or go high from the range ?

     

  • arbitary32arbitary32 Member Posts: 3 Contributor I

    @rfuentealba is it possible to create the the dashboard in line graph and display the dashboard thorugh smartphone whenever  smartphone is connected with internet and also share the notificatio alert whenever temprature drops or go high from the range ?

  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member, University Professor Posts: 568 Unicorn
    Hi @arbitary32,

    It can be possible to do it the other way round. Build your dashboard and make it get the data from the server. Don't see it unfeasible.
  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member, University Professor Posts: 568 Unicorn
    Hi @arbitary32,

    Regarding notifications: if your endpoint saves the data, applies the proper model to evaluate it, and you have the Python Extension, you can do whatever you want: send push notifications, SMS, emails or even publish a random sentence on Facebook based on your values :P There is no push notification functionality inside RapidMiner Server, but you can use another one like Twilio.
Sign In or Register to comment.