Reading External Data Sources | Beginning Azure ML | Part 2

Key Takeaways

all right hello internet welcome back to our web tutorial on azure ml so last time i showed you guys how to create an experiment and import data from a local file and today i'm actually going to teach you how to read data from an external source so last time i showed you how to import the titanic data set right so we just drag it in so that was an example of reading from a static data store something that you saved on your local hard drive but if you remember correctly we got that data from an external website right so you can actually have azure ml read that directly from the link itself so go to data input and output and in that you'll find a module called the reader module all right so just drag that in the reader module is actually pretty flexible for what it can do and the amount of inputs it will take in so you'll take a bunch of inputs from a variety of sources so an external link is represented as this http link right here so you select http and i'm just going to go to an external website that hosts a number of data sets right so the titanic just happens to be one of these data sets so i'm just going to copy the csv link so i'll go back to my reader module and just paste it into the url once i've selected http and yes it is a csv file and yes the csv does have headers in it after you've done all that this actually doesn't contain any data inside of it yet still just an empty module you will actually have to run the model down here for it to actually read and parse the data you don't have to do this right now but it's just good practice to check to see if your data is actually there before you start doing stuff with your data so i'm going to hit the run and what's going to happen now is a timer is going to appear in the module that means the cloud is working and it's processing and it's parsing your data from the external source so you can tell that it's done by the check mark next to the module and you can actually see what's inside a quick preview of what's inside ah okay so the data got parsed into the module that's really cool and that was how to read from an external link now this is really cool right because if tomorrow the data change you can just hit run again tomorrow and it will read tomorrow's data so if you have ever changing data if your data is dynamic and it's continuing to change you can continue to update your model and it won't affect the rest of your experiment it will continue to flow through the rest of the experiment in the normal workflows as if you know it was the same data now it can support a bunch of other things too so that was the http so let's just select this module ctrl c to copy it and then ctrl v to paste it now we have another reader module right so this one was reading the data from the external link now this will actually support a bunch of other things as well so for example you can do a hive query all right so it supports hadoop a power query or you can read from a sql as your table so you can literally start typing in sql right so i'm just going to do a quick demonstration so select column from table where date like today right so that will get you for example the sales data from today now what if you want to compare it you know with yesterday's data all right so just copy and paste it and then just say yesterday right so now you have two back-to-back data files reading two separate sql tables i would recommend adding a caption to uh your reader module just so that it's very clear so just double click on it and label what it is say today's merchandise sales data right and then an arrow should appear that you can close or open and it will tell you under what the label is and that you could do this for all modules not just for the reader here's another quick example of what the reader can do so here is a reader that's reading from a titanic data set that we've backed up to the azure blob storage so it is backed up securely in the cloud so if you want to share this securely or back it up securely and not have it be an open http link you would just upload it to the azure blob storage and then read it from the azure blob storage and at any time you have a data set that's populated via the reader module you can simply just click save as data set on the node and it will actually just show up over in the data sets area and that concludes the tour on how to read data from external source as well as the reader module join us next time where i will show you how to do data exploration and data visualization using the tools present within azure ml if you like what you just saw subscribe to our channel or leave us a comment let us know if there's a topic you want us to cover and be sure to check us out at datasciencedojo.com until next time