Watson training data format - json

I want Watson to train on certain data the user will provide in my web app - the data will be posted through forms.
My question is - which service in IBM Cloud matches this the best? I've tried Discovery, but it doesn't seem to be the best match for my request, first of all it does not want to accept .json or Excel formatted files, which seems like a red flag to me (concerning what I am looking for).
My ultimate goal is for Watson to learn the patterns and ultimately start providing suggestions for the user.
My data I give to Watson would look like this, .json format:
{
"songName" : "Beyond the sea",
"artist" : "Bobby Darin",
"genre" : "jazz"
}
Thank you in advance.
I've setup my IBM Cloud, enabled Discovery as a service, attempted to upload .json and excel files, from which both have been rejected.
I expected Watson to process the provided structured data, find patterns, and provide intelligent suggestions.

If you are planning to build and train a model, then Watson Machine Learning service on IBM Cloud is what you are looking for. After you build and train a model, you deploy it and put it into production, so you can pass data to the model and get scoring data, also known as predictions, back.
You can access your deployment through an API endpoint, using the Watson Machine Learning Python client, CLI, or REST API (manually or in an application) to analyze data.
To know more about deploying your model, refer this the documentation here
Also, there is a code pattern that serves as an example of how to build a Machine Learning recommendation engine using Jupyter notebooks on Watson Studio

Related

Amazon MWS and Microsoft Access for a DB Layman?

I have some experience in MS Access, but mostly only as an offline DB tool.
I have begun working with both Seller and Vendor Central at my new company, and am in charge of scrubbing the vast amount of data for trends and whatnot. At the moment our company is solely relying on exporting reports from Seller Central directly, and cross referencing documents. I was hoping to get us started with a rudimentary database hooked into Seller directly. Our company already has a MWS Developer ID, and I see an MWS Access Key and whatnot.
I'm surprised to not finding any resources as to how I should actually connect MWS to Access. I feel confident that I can find some success by dabbling with the API once I get it connected, but I can't actually find any references on how to actually establish that connection.
Any resources you guys can forward me? Maybe I'm searching for the wrong terms. Everything I search just comes up with data service companies advertising their tools.
Well, the interface to AWS is going to be web service based. And access unfortantly does not have a built in web services interface.
So, your choices are:
Write some VBA code to hit/use/consume AWS web services. Web services are just that -a web API. (likly REST services. REST is just a fancy term that you have to type in a given URL.
So, what you looking to search for?
How can I consume web based data in Access.
Say this answer on SO
Making a SOAP request from Access 2007
The main issue is that Access does not have really good tools for consuming web data.
However, most web front "store" applcations tend to have a user area in which you can export the daily sales or data say to csv. You now can import that data into Access (or Excel).
And they often have a report area - you can generate a report, and then download again in some format like xml or csv (and again, import into Access or Excel).
If you don't want to have to maually import the data?
Then you have to code out web requests. And that can be painful.
This unfortantly means you can use say a linked table (ODBC) like you can for Acces say to some database.
So, you can start to write web interface code (it will be SOAP or REST.
Believe it or not, there was a SOAP add-in tool kit for Access 2003. But, no one used it, so they dropped it. (of course now 17 years later -gee, a truckload of people GET IT - and now see the need to consume web data!
So, you question and what to learn about?
You asking how does one consume web services.
Well, using a tool designed to work with web services helps a lot. (that's why I suggest Visual Studio and .net). If they have a WSDL for you? Then you can point Visual Studio at the web (WSDL), and it will crank out a set of "methods" and properites for you. (it will create a class. But then again, did you use and write class objects in VBA? (it does support you creating classes. But the SOAP tool kit (no longer avaiable) would write this code for you!
So, if you want to go beyond their built-in repoting tools (that let you export + download the data in some format like csv for use with Access or Excel)?
Then you have to write writing code to make web calls.
This is not a lot different in the past. If you wanted some data from the accounting system? Well, you can/could/usually do some export with the accounting package to spit out a csv file of some sort. You then import into Access.
However, if you had better skills, you might link up to the database from Access, using ODBC and then write some SQL queries against that data. So, it really comes down to skill level here. Some could not be bothered to learn say SQL and a query. So, they just export the data out of accounting, and then import into access.
The problem is now you can't link to that web site, and use SQL queries of data. You have to use web service calls. (at least if you want to make some of this process automatic).
So, you might be just fine by exporting data/files from the AWS services, and then just import into Excel or Access. As such, you not writing any code, and you just use the Access GUI to import data.
But, some want to just hit a button in Access, and see all the orders and sales from today - and have Access pull that data from the web site with one click.
For some simple data pulls? You could make a web call from Access. But for complex web interfaces? Then you need to use tools that support web interfacing (say like Visual Studio .net).
For a simple data pull? I'll use VBA and MSXML.
But, if the parameters and data call is complex? Then I write it in .net, and THEN expose that code as a consuming library to MS-Access.
So, once you signed up for AWS and what ever web services? Then they will supply you with the web calls, and documentation. You then are free to use your programming tools of choice to interface. But, this can be quite a bit of work. So, you might use VBA, but .net is much better for this type of work. (and it also a lot more difficult to code out).
As a developer who has done this, I would write a "sync" program that connects to MWS, pulls back your data, and then inserts that into MS Access. In my case, it was a C# .NET Core app with SQL Server and I used the available MWS SDK that Amazon provides for free to handle all the API calls to MWS. You can create a schedule so your app pulls the data on an interval, or make it manual where you push a button to sync it into your system.
Of course you can use Java or PHP instead of C#, or you can roll your own MWS API calls. Or like you mention there are several third party vendors that have out-of-the-box ready solutions.
I haven't used MS Access in 20 years or so, so I'm not sure about calling MWS directly. I would gather it could be done, but is probably too much work, but I could be wrong. A .NET app can insert into MS Access, no problem, but also handle the HTTP calls to MWS for you.

Recommendation for BigQuery Reporting/BI Tool

I work for web hosting company looking to integrate different data sources with BigQuery but the question now is what would be an ideal reporting/BI tool to get the data from BigQuery so proper/fast/easy retrieval/analysis/ reporting can be done with it.
I'm looking into the options suggested by google here: https://cloud.google.com/bigquery/partners/ but I was wondering if someone out there has possibly a more hands-on experience that could make a recommendation.
the company works with a mysql based billing system (with client, support, service data) which is the main source of info, along with other chat, cms and inhouse-developed systems that provide other sources of information that allow to maintain the web infrastructure where the business depends on.
Thank you.
It's really hard to answer this. Depends on the personnel you have at hand.
We are doing for idea validation mostly Data Studio.
Some personnel knows Tableau, but once you are out from GCP, all become a slow process, queries and interface updates in 30-60 seconds, as they all relay and store on their own the data.
We have wired some data to ElasticSearch as well, and we use Kibana.
But once it's all validated, we consolidated into our own Dashboards the reports. Mainly because we are mostly developers and can do the programming. If you have a data analyist or data scientist with their own tools, let them use what they are comfortable with.
Always do iteration and versioning, you as a developer should be driven by a good product manager who tells exactly what charts to build out.

How to make two json file configurations for different Firebase projects?

We have two Firebase project, one for developing another production project. We use cloud functions. In one cloud functions, you need to use service-account-credentials.json. The problem is how can I make this function take data from service-account-credentials-dev.json when it proceeds to the development project, and when on production, then from service-account-credentials-prod.json?
I know about the environment, but as I understand, this feature does not allow you to download the json file for a particular project.
I found the answer to my question here. Doug Stevenson wrote "There is not. You could write your own program to read the json and convert that into a series of firebase CLI commands that get the job done"

OpsCenter backend API for generating custom reports

I would like to know if we have any specific api's in OpsCenter which would help in generating custom reports.
The use case is:
I am consolidating reports from different tools like App Dynamics,OpsCenter and others into one csv or excel sheet for which i would require any opscenter api which would allow me to do a backend call which would pull the required info for me from OpsCenter and then i can do some additional stuff and consolidate it along with other data in a csv or excel in an automated way.
Any api which would help me with backend Opscenter calls for reports data would be helpful.
The OpsCenter API is fully documented at http://docs.datastax.com/en/latest-opscenter/opsc/opscApi_g.html
If the kinds of reports you're looking to generate involve metric data, you'll be particularly interested in the new-metrics endpoint: http://docs.datastax.com/en/opscenter/6.0/api/docs/metrics.html#method-get-new-metrics
Otherwise you'll just have to wade through it to find what you're looking for. But the front-end is all driven through those api's, so anything you see in the web-interface can be replicated via the documented apis.

Integrating CSV person data on Google Maps through Bluemix?

I am working on a project with a large CSV file that containes the location and movement of users. I would like to place this on a custom map in Google Maps via bluemix and use Bluemix Services to explore the data.
The primary goals are:
Getting the CSV data on the custom Google map. When running, the data should progress in time and show the movement of users.
Making the CSV points cluster for UX. (so that points that are near each other would stack together)
My primary question is how to get started on this. Do you reccommend i work on this locally and then connect Bluemix to my project or can i create all of this in Bluemix. I would much prefer the last option if possible.
If you have any suggestions to Watson Services or other Bluemix Features that may improve the app this is also greatly appreciated
Thank you for your time.
Ps. I realize Google Maps integrates best with Java Script. Do you recommend converting the CSV to Json when working with Bluemix?
This is a quite broad question and maybe not well suited for stackoverflow (stackoverflow is not a discussion forum, it is a competition of what answer is the most accurate one for a very strict question) so maybe this is a question for https://developer.ibm.com/answers/ instead.
That said, bluemix works well with js, java, ruby, python, go, php, etc (probably js and java better than the others) so I'd go in this direction. Also, I think you should investigate bluemix geospatial analytics (https://console.ng.bluemix.net/catalog/services/geospatial-analytics/) for your application.
For data storage, I suggest you to take a look on cloudant (https://console.ng.bluemix.net/catalog/services/cloudant-nosql-db/) which is a very popular option in bluemix and suits well for most cloud apps. If you want to take a more traditional approach, you can also consider a relational db such as DB2 (https://console.ng.bluemix.net/catalog/services/ibm-db2-on-cloud/)