Contents


Build your first social media dashboard in minutes with Node-RED and Db2 for i

Unleash the power of cloud services and Db2 for i with open source solutions, Node-RED and Node.js

Comments

How to easily and rapidly create business values from open source technology, cloud services, and your data on IBM i? Make your idea a reality by prototyping like a master using open source technologies, Node-RED, and Node.js in conjunction with IBM® Db2® for i!

In this article, we implement (with almost no lines of code and in a few minutes) a simple social media dashboard, including sentiment analysis, natural language processing (NLP) and business data enrichment. Stop talking and simply do it!

Figure 1. Social media dashboards using Node-RED, IBM Watson, and Db2 for i

Social listening is the process of monitoring the conversations and mentions of your brands online, and is valuable for all organizations, marketing teams, customer support teams, business development, and competition surveys, especially when used and mixed with business data.

This process allows you to get real-time insights in many use cases. Some of them are listed below.

  • How your brand is mentioned in the social media
  • How you can adjust marketing campaigns based on metrics such as impressions and engagements
  • How you can have more information about your client segmentation
  • How to react to customer complaints in real time

The objective here is to highlight the capabilities of open source technologies available on IBM i (integrated with Db2 for i), and illustrate how to get started easily with digital transformation, cognitive computing, APIs, and hybrid cloud.

Node-RED is an open source flow-based coding tool coming from Internet of Things (IoT), brought by IBM Emerging Technology, proposing a large number of nodes from the node-red.org community. Besides IoT, it is ideal for building micro-services based architecture solutions, prototyping, or creating and consuming cloud or on-premises APIs with no lines of code.

Need more information about Node-RED on IBM i and how to get started?

Refer IBM developerWorks article, Running Node-RED on IBM i: Installation and first flow.

This article explains how to implement real-time sentiment analysis for particular hashtags and write this information in your Db2 database, involving a Twitter node, optionally an IBM Watson™ Language Translation node, a sentiment analysis node, and a Db2 for i node to enrich business data. It also describes how to visualize data stored in Db2 for i and create a simple social media dashboard web application using dashboard nodes and Db2 for i nodes.

The optional Watson Language Translation node will identify languages and translate tweets in English on the fly before sentiment analysis.

What you'll need

To write this social media dashboard, you need the following minimum prerequisites:

  • 5733OPS option 10 that is, Node.js v6, Node-RED installed, and a web browser. Refer to the "Running Node-RED on IBM i: Installation and first flow" article to view all the detailed prerequisites.
  • A decent Db2 for i SQL editor connected to an IBM i 7.2 system and below. You can use IBM i Access Client Solutions and the embedded SQL editor.
  • A Twitter account, and a connection to the Internet from IBM i. Optionally, you can simulate tweets using the test data provided in this article.
  • Optional: An IBM Cloud (former IBM Bluemix®) account for using IBM Watson and a Watson Language Translation node.
    You can register for a free trial or freemium account or log in to IBM Cloud if you already have an account.

Start Node-RED

Refer to this article if you want to know more about Node-RED and how to get started on IBM i.

Start Node-RED and connect to it for example, using the default URL: http://<IBMI-IP>:1880 where <IBM-IP> is the IP address of your IBM i system.

It is a good practice to set authentication credentials to your Node-RED and switch from HTTP to HTTPS.

Create a basic tweets table in Db2 for i and import sample data

Let us consider an example where we store only a subset of the information contained in the incoming tweets, but it is good to notice that Twitter APIs return a lot more attributes in a JSON object for each tweet, as documented on the Twitter Developer Platform website.

Also note that the Node-RED Twitter node returns all this information in JSON, and we will parse and then store this data in a standard Db2 table for the sake of simplicity. However, it is better to use Db2 for i for directly storing the incoming tweets in JSON in a CLOB column.

You can then retrieve information about tweets with the powerful JSON_TABLE functions available on Db2 for i.

  1. Create a simple Tweets table for storing our Twitter activity. For that purpose, open a SQL editor, for example IBM i Access Client Solutions and run the following SQL statements for creating a sentiment library, a Tweets table, and a simple trigger for that table.
    	CREATE collection SENTIMENT;
    	CREATE OR REPLACE TABLE SENTIMENT.TWEETS (
    	    id SMALLINT NOT NULL
    	        GENERATED ALWAYS AS IDENTITY
    	        (START WITH 1
    	        INCREMENT BY 1
    	        CYCLE),
    	    TWEET VARCHAR(5000),
    	    SCREEN_NAME VARCHAR(255),
    	    LOCATION VARCHAR(255),
    	    SENTIMENT_SCORE INT,
    	    SENTIMENT_GROUP VARCHAR(255),
    	    TIMESTP TIMESTAMP,
    	    TIMESTAMP_MS VARCHAR(255),
    	    primary key (id)
    	);
    	CREATE OR REPLACE TRIGGER SENTIMENT.Tweets_sentiment
    	  BEFORE INSERT OR UPDATE ON SENTIMENT.Tweets
    	  REFERENCING NEW AS n OLD AS o
    	  FOR EACH ROW MODE Db2ROW
    	  BEGIN
    	    SET n.SENTIMENT_GROUP = CASE WHEN n.SENTIMENT_SCORE > 0 THEN 'positive'
    		                             WHEN n.SENTIMENT_SCORE = 0   THEN 'neutral'
    	                                 WHEN n.SENTIMENT_SCORE < 0  THEN 'negative'      
    	    END;    
    	 END;

    If you want to use an existing library on your system, just copy the CREATE TABLE and CREATE TRIGGER sections above, and start journaling the Tweets table.

    The Tweets_sentiment trigger will just update the inserted or updated record with a SENTIMENT_GROUP (positive, negative, neutral) depending on the SENTIMENT_SCORE value.

  2. Let's populate this table with sample data (with 12 public tweets). Later in this article, our Twitter node will populate the Tweets table for us.
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('What an impressive panel at our ask the experts during our conference! #ibmi #cshk2017 
    https://t.co/StgAsBPFXv','MBuurRasmussen','nolocation',3,CURRENT TIMESTAMP,1508850681582);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('Thanks, @chris_lalevee for this bodacious article about Node-RED on #IBMi!  #IBMiOSS 
    https://t.co/yursu7d2GJ','IBMJesseG','Rochester, MI',2,CURRENT TIMESTAMP,1508850903650);
    
    INSERT INTO 
    SENTIMENT.TWEETS (TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES ('RT @IBMi: Thank you, for teaching us all a valuable lesson, 
    https://t.co/idje2J3','IBMiFan','United States of America',6,CURRENT TIMESTAMP, 1508851116842);
    
    INSERT INTO 
    SENTIMENT.TWEETS (TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('RT #IBMi Opensource tools https://t.co/s8sjeu','ClubCEO','Texas, USA',0,
    CURRENT TIMESTAMP,1508851142085);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('Excellent article about #ibmi & Db2 here https://t.co/s8sjeu!!','MariaMe6452','Dublin, 
    Ireland',3,CURRENT TIMESTAMP, 1508851159475);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('Meet the next generation of #IBMi and #IBMAIX with these #IBMFreshFaces featured in @IBMPowerMag:  
    https://t.co/PbpnGpypnE','IBMPowerSystems','Sugarland, TX',3,CURRENT TIMESTAMP, 1508851165417);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('RT @ibm: Our Enterprise #IBMi Modernization White Paper helps you understand how #Nodejs can 
    attract new development talent! ','3E8TestBot','Lost in Space',5,CURRENT TIMESTAMP, 1508851192995);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('RT @kadler_ibm: Good news, everyone! Come see me at @zendcon and we can talk all things 
    #IBMiOSS and #PHP on #IBMi. I might even ha ','phpc','Freenode IRC',5,CURRENT TIMESTAMP, 1508851202542);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('RT @HelpSystemsMN: What are #IBMi special authorities & who should have them? 
    Security expert @DaljitSingh explains > ','J_Buck51','Kenosha, WI',0,CURRENT TIMESTAMP,1508851319619);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('RT @chris_lalevee: Install @NodeRED on #IBMi (https://t.co/DHlifEAEuo) 
    and test Db2 for i Node !! https://t.co/KKu5Ji5ihW','MarolleauBenoit','Montpellier, France',0,
    CURRENT TIMESTAMP,1508851419619);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('Looking forward to meeting #IBMi customers next week at the Jack Henry Conference in Nashville. 
    Stop by booth 208 https://t.co/gZs4MXNsO8','petem59','Connecticut, USA',0,CURRENT TIMESTAMP,1508851544257);
    
    INSERT INTO 
    SENTIMENT.TWEETS(TWEET,SCREEN_NAME,LOCATION,SENTIMENT_SCORE,TIMESTP,TIMESTAMP_MS) 
    VALUES('RT #IBMi #db2 #XML Des soucis de gestion dates/heures avec XMLTABLE ? Allez plus loin avec les 
    fonctions XQuery/XPath  ','PhBourgeoisIBM','Paris, France',0,CURRENT TIMESTAMP,1508851704005);

Install the Db2 for i node

First, we need to install the Db2 for i node in our Node-RED palette. You can do it by using the command line as explained on the Db2 for i node home page or by using the Node-RED graphical interface as illustrated below.

  1. In Node-RED, click the button at the upper-right corner and click Manage palette.
    Figure 2. Add nodes to the Node-RED palette
  2. On the Install tab, search for db2, and then search for node-red-contrib-db2-for-i, and click Install. Wait for the installation confirmation dialog box to be displayed, and then restart Node-RED.
    Figure 3. Add Db2 for i “node-red-contrib-db2-for-i’ nodes to the Node-RED palette
  3. Browse the node palette on the left side and notice that the Db2 for i node is available in the Storage category.
    Figure 4. Db2 for i node on Node-RED

Test your first flow with the Db2 for i node

Let's test the node with a simple query. Drag the Db2 for i node to a new flow, as shown in the following figure. Then, add SQL Query as the inject node and msg.payload as the debug node.

Figure 5. Simple Node-RED flow: Input Inject node – Db2 for i node – Output Debug node
  1. Configure the Inject node with the settings and SQL query as shown in the following figure.
    Figure 6. Inject node setup with a SQL query
  2. Configure the Db2 for i node with the settings as shown in the following figure.
    Figure 7. Db2 for i node settings - adding a Db2 connection
  3. Set up a configuration node by specifying a connection name and, optionally, a user name/password if you don't want to use the current user profile running Node-RED, and you want to specify a particular user profile for connecting your database. Then click Add and Done.
    Figure 8. Configure a “Db2 for i Config” node with optional credentials
  4. Test this simple flow by clicking Deploy and then click the Inject node button on the left side of your flow as shown in the following figure.
    Figure 9. First Node-RED flow injecting a SQL query to a Db2 for i database

Look into the results (JSON array of rows) of the SQL query in the Debug panel.

Figure 10. Db2 for i JSON result set in the Node-RED debug panel

At this step, if you see no result and if you are sure that your table exists and is not empty, check if the user profile used has read access (with appropriate SQL grant or IBM i authority).

You've just written your first Node-RED flow using Db2 for i !

Call Twitter APIs and feed your Db2 for i database with tweets

The objective here is to make your IBM i capable of listening to tweets and analyzing sentiment on the fly before inserting them into your Db2 for i database.

To do so, you just have to manually drag and configure four nodes in your Node-RED flow editor.

To do this quickly, import an existing flow into your current Node-RED flow:

  1. Click the menu at the upper-right corner and click Import->Clipboard.
    Figure 11. Flow import from clipboard in Node-RED
  2. Extract the .json file from the 1_ibmi_twitter_listening.zip file (also available in the Downloadable resources section) and copy the flow from the .json file and click current flow. Then, click Import.

    You should get the following result.

    Figure 12. First Twitter listening Node-RED flow with sentiment analysis and insertion in Db2 for i
  3. Configure your nodes.
    • Twitter node: Specify your Twitter user account and hashtag to perform a search. If you don't have a Twitter account, create one or skip this step.
    • Db2 for i node: Specify connection information as explained in the previous step.
    • Function node (PrepareForInsert): Control the schema and table name reflect in your environment in the JavaScript code by double-clicking on that node and modify the JavaScript function if necessary.
  4. Click Deploy and wait for Twitter activity. You can optionally add a Debug node in your flow to see the Twitter activity in the Debug view.
  5. Verify that your Tweets table is being populated by querying your table occasionally, either from your favorite SQL editor or from a Node-RED flow.

Optional: On the fly tweets translation using IBM Watson services

We can enrich incoming tweets with IBM Watson services, for example, to translate the incoming tweets into English on the fly before performing sentiment analysis. This part is optional, so you can possibly go to the dashboard creation section directly.

  1. Instantiate a Watson Language Translator service from the IBM Cloud catalog, and get the generated credentials (user name, password, and endpoint). You'll need them when configuring the associated node in Node-RED.
    Figure 13. IBM Watson Language Translator service in the IBM Cloud catalog
    Figure 14. IBM Watson Language Translator service credentials
  2. Copy a sample flow to your existing Twitter flow and configure your nodes with your credentials.
    1. Extract the .json file from the 2_ibmi_watson_translator_subflow.zip file (also available in the Downloadable resources section) and copy the flow from the .json file to your clipboard.
      Before proceeding with the next step, you need to have the IBM Watson nodes installed in your Node-RED palette. If it is not the case, install the node-red-node-watson nodes using the procedure described in the Install the Db2 for i node section. The name of the package to install in the palette is node-red-node-watson.
    2. In your Node-RED editor in the menu at the upper-right corner, click Import -> Clipboard. Paste the code, and click New Flow and then click Import.
      This will import a Node-RED subflow. Now, add an additional node in your palette on the left side of your editor as shown in the figure below.
      A Node-RED subflow is appropriate when you want to make a portion of a flow reusable for other flows, without having to duplicate it. Click Deploy for saving and deploying the imported subflow to your Node-RED workspace.
      Figure 15. Creation of an IBM Watson-based subflow in Node-RED
    3. Double-click Watson Translation to English to edit the newly created subflow, including the IBM Watson nodes.
      Figure 16. Details of an IBM Watson-based subflow for language identification and translation

      This subflow first identities the languages used in the input message, and translates it if different from English, using IBM Watson services and APIs.

    4. Configure the two Watson nodes (by specifying the user name, password, and endpoint if needed), each pointing to the same IBM Watson service you have just instantiated on IBM Cloud, but each using a different API function (language identity and language translation).
    5. Test the subflow with a test flow, for example, using the manual injection nodes, as illustrated in the following figure.
      Figure 17. Test of your first IBM Watson Language Translator Node-RED flow on IBM i


      Thanks to IBM Watson, Your IBM i is now translating French, Spanish, and many other languages, into English!

  3. Integrate this flow within your existing Twitter flow by connecting it graphically, as show in the following figure.
    Figure 18. End-to-end Node-RED flow with social media listening, IBM Watson translation, and Db2 for i
  4. Test the result either by injecting an input test payload (using the inject node) or by waiting for a new incoming tweet.

    Your IBM i is now doing Social Media listening and translating social media data on the fly with IBM Watson before inserting them into your Db2 for i database!

Visualize your Db2 for i and social media data with a simple dashboard

In this section, we are, without coding, going to create a social media dashboard, mixing your business data on Db2 for i with data coming from the IBM Cloud and IoT.

  1. Install the dashboard nodes.

    Many Node-RED exist for generating UI and graphs. The node-red-dashboard is one of my favorite.
    From your Node-RED editor, in the upper-right menu, click Manage palette, and search for node-red-dashboard. Then, click Install and restart Node-RED.

    Figure 19. Installation of “node-red-dashboard” in Node-RED
  2. Extract the .json file from the 3_ibmi_simple_dashboard.zip file (also available in the Downloadable resources section) and import the code from the .json file to the Node-RED flow already created.
  3. Notice that a new flow as shown in the following figure is displayed.
    Figure 20. Simple dashboard nodes with Twitter and Db2 for i
  4. Configure each Db2 node with your connection information.
  5. Deploy and access your dashboard. The refresh rate in the just imported code is set to 10 minutes. So, if you don't want to wait, go to your flow editor and force a dashboard refresh by clicking each inject nodes on the left side of your flow. The dashboard URL is available at the right side of the dashboard panel.
    Figure 21. Node-RED dashboard layout panel and URL

You have just created a simple social media dashboard, with a few graphs populated with data residing on your Db2 for i database!

Figure 22. Social media dashboard on IBM i with Node-RED, Twitter APIs, and Db2 for i

Summary

"Fail fast, succeed faster"

IBM i is a private cloud on its own, and when used with open source technology in conjunction with Db2 for i, you can perform great things quickly and bring innovation for your business.

In this article, we demonstrated that Node-RED and Node.js on IBM i can be a real accelerator for API integration and prototyping, allowing you to play with nodes and program the whole world by a simple drag-and-drop action! We also had an overview of the power of IBM Watson and its cognitive computing APIs for augmenting all our solutions on IBM i.

References


Downloadable resources


Comments

Sign in or register to add and subscribe to comments.

static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=IBM i
ArticleID=1052914
ArticleTitle=Build your first social media dashboard in minutes with Node-RED and Db2 for i
publish-date=11282017