Qlikview and Google BigQuery – Data Visualization for Big Data

google launched its bigquery cloud service in may to support interactive analysis of massive datasets up to billions of rows. shortly after this launch qliktech demonstrated a qlikview google bigquery application that provided data visualization using bigquery as backend. this post is about how qlikview and google bigquery can be integrated to provide easy-to-use data analytics applications for business users who work on large datasets.

qlikview and google bigquery

qlikview has two capabilities depending on the needs and the volume of the data:

qlikview bigquery connector: this add-on is written in .net – thus it requires microsoft .net 4 framework to be installed on your computer -, it loads the data into the in-memory data model and various view types (table, barchart, etc) can be then created on the fly to visualize the data or its subset.

qlikview bigquery extension object: in case of a huge volume of data not all the data can be loaded into memory. qlikview bigquery extension object provides a web-based solution, it is built upon google javascript api . users can navigate using the extension object and get only the relevant portion of the data from bigquery.

preparing the dataset in google bigquery

before we start working with qlikview bigquery solutions, we need to create a dataset in google bigquery. we are going to use apple marketdata donwloaded from finance.yahoo.com site in csv format.

first we need to create a dataset called apple using google bigquery browser tool :

then we need to create a table called marketdata:

the next step is to upload the csv file into the  table – you need to use chrome browser, as internet explorer does not work for file upload as of the writing of this post. the schema that was used is {date:string, open:float, high:float, low:float, close:float, volume:integer, adjclose: float}- just to demonstrate the google bigquery is capable of handling various data types:

finally we can run a simple sql query to validate that the data has been successfully uploaded:

so far so good, we have the data loaded into google bigquery

qlikview bigquery connector

we need to have qlikview installed, in my test i used qlikview personal edition that can be dowloaded for free from qliktech website. then we need to download qlikview bigquery connector  from qlikview market .

once qlikview bigquery connector is installed, it appears in a similar way as any other connectors (just like odbc or ole db). go to edit script and then choose bigquery as database:

once we click on connect, an authorization window pops up on the screen – google bigquery relies on oauth2.0, thus we need to have oauth2.0 client id and client secret. the client id and client secret can be created using google api console . select ‘installed application’ and ‘other’ options.

in qlikview we need to authenticate ourself using the client id and client secret:

after authentication the next step is to define the select statement that will be used to load the data from google bigquery into memory:

when we click on ok button the data is being fetched into qlikview in-memory data model (in our case it is 7,000+ lines):

we can start processing and visualizing the data within qlikview. first we are going to create a table view by right click and then selecting new object sheet:

let us then define another visualization object, a line chart:

and then a barchart – so we will get the following dashboard to present the data that was loaded from google bigquery backend into qlikview in-memory column store:

qlikview bigquery extension object

as said before, not necessarily all the data can  fit into the memory – even if qlikview is very strong at compressing data, we are talking about massive datasets, aren’t we – that is what big data is all about. in this case qlikview bigquery extension object comes to the rescue. we need to download it from qlikview market and install it.

as it is a web-based solution using javascript (google javascript api), we’d better have a google client id and client secret for web applications, we can create it in the same way as described above for qlikview connector. the ‘javascript origins’ attribute needs to be modified to http://qlikview.

then we need to turn on webview in qlikview:

now we are ready to create a new visualization object by right click and selecting new sheet object:

select extenstion objects, bigquery and drag it onto the qlikview window

we need to define the project, dataset and table – bighadoop, apple, marketdata respectively:

then we have to define the visualization type (table in this case) and the select statement  to fetch the data. please, note that we limited the data to 100 lines using ‘select date, open, close, high, low from apple.marketdata limit 100; sql statement:

we can define various visualization objects, similarly to the bigquery connector scenario:

qlikview on mobile

qliktech promotes a unified approach for delivering bi solutions for different platforms based on html5, no need for additional layers to support data analytics and visualization on mobile devices. qlikview server is capable of recognizing mobile browsers and supports touch-screen functionalities.

reference: http://www.slidshare.net/qlikview_uk/qlikview-on-mobile

conclusion

qlikview has gained significant popularity among bi tools, gartner positioned qliktech in the  leaders zone of the business inteligent platforms magic quadrant in 2012. it provides highly interactive, easy-to-use graphical user interface for business users and the technology partnership with google to provide seamless integration with bigquery can just  further strengthen its position.