Google Prediction Connector For QlikView & Qlik Sense

Modified on 2016/02/15 10:09 by chrisb Categorized as connector, Text Analytics
{toc}

The Google Prediction Connector allows you to [Creating-A-Training-Data-Set-And-Google-Predict-Model-Using-QlikView-And-QVSource|train your own prediction models] (both regression models and categorisation models) and then run data in your QlikView applications through these prediction models in a structured and efficient way.

Google give a few examples on their website of how this might be useful:
* Document and Email Classification
* Recommendation Systems
* Spam Detection
* Language Detection
* Upsell Opportunity Analysis
* Diagnostics
* Suspicious Activity Detection
* Churn Analysis
* Sentiment Analysis

And you can see a guide on using this connector [Creating-A-Training-Data-Set-And-Google-Predict-Model-Using-QlikView-And-QVSource|here] which uses the last example above of sentiment analysis.

{toc}

==Getting Set up==
===Create An API Project===
You need to first set up your own Google API Application and Cloud Storage Area before you can start using this Connector. Please follow the steps below. Once this is set up you can follow [Creating-A-Training-Data-Set-And-Google-Predict-Model-Using-QlikView-And-QVSource|this guide] on using this connector.

The first step is to go to the [https://code.google.com/apis/console|Google API Console] and create a new project:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Create Project.png]

Give the project a name:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Name Project.png]

===Set Up API Access===
On the Services tab, turn on access to the Google Prediction API:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Google Prediction Activated.png]

And Google Cloud Storage:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Cloud Storage Activated.png]

===Enable Billing===
Next, enable billing. You will need to follow a few steps here to set up payment information.

(((You can find pricing information for the Google Prediction API [https://developers.google.com/prediction/docs/pricing|here], but at time of writing it is $10/month - which included 10,000 predictions, and then $0.50 per 1000 predictions thereafter.)))

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Enable Billing.png]

Make a note of the project number as you will need this when using the connector:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Train Model - Project.png]

===Create An OAuth 2.0 ID===
On the API Access tab, create a new OAuth 2.0 ID:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Create OAuth2 Client ID.png]

Give it a name:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Create OAuth2 Client ID - Step 1.png]

And configure it as an installed application:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Create OAuth2 Client ID - Step 2.png]

And then note the client id and client secret as you will also need these when using the connector.

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Create OAuth2 Client ID - Client ID And Secret.png]

===Set Up Cloud Storage===
Next, select the Google Cloud Storage option:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Google Cloud Manager.png]

Create a new bucket:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}New Bucket.png]

And give it a name:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Bucket Name.png]

It is here where you will upload your training data, after which you should note the path to the file (bucketname/filename) - but you can wait to read [Creating-A-Training-Data-Set-And-Google-Predict-Model-Using-QlikView-And-QVSource|this guide] before doing anything further here.

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}Select Training Data.png]

==Summary Checklist==
If you have completed the above stages correctly you should have:
* Set up a new Google API Project, with billing enabled and access to the Google Prediction API and Google Cloud Storage API.
* Noted the '''project number''' of this.
* Set up a new OAuth 2.0 Client and noted the '''client id''' and '''client secret'''.
* Set up a bucket in Google Cloud Storage.

These three values (in bold above) are entered into the Connector as highlighted below:

[image||{UP(Google-Prediction-API-Connector-For-QlikView)}enter config.png]

With all this in place you are ready to [Creating-A-Training-Data-Set-And-Google-Predict-Model-Using-QlikView-And-QVSource|create a training data set and start using the connector].

==Change Log==
_0.8.3 - 15/02/16__
* Minor internal refactoring.

__0.8.2 - 02/02/16__
* Removed dependency on Microsoft.VisualBasic assembly.

__0.8.1 - 26/03/14__
* Minor performance improvements.
* Requests to the Google Predict API should now also be logged to the application's API Log.
* Fixed BUG with client ID not being set before authenticating.

__0.8.0 - 17/09/13__
* Initial version.
Meta Keywords:
Meta Description:
Change Comment:
(QVSource works with Qlik Sense as well as QlikView - See this page for notes.)
QVSource: The QlikView & Qlik Sense API Connector (Docs)

Navigation¶

Google Prediction Connector For QlikView & Qlik Sense
