In my “Reinventing online reviews” article I designed a unique method for categorizing and viewing online reviews. Since then, I have built a working prototype application for this design. I also added integration capabilities with Twitter to visualize sentiment for a given topic.

Hexagon Geometry

One of the challegening parts of this proejct was translating the vision I had of a hexagon structure into a tangiable web application.

To achieve my grid of hexagons design I needed to learn how to arrange the hexagons into a specfic manner. A blog post from Amit Patel of Red Blog Games on hexagonal design provided alot of the base information on drawing and arranging hexagons. The image below was the foundation for what I built.

Within my code my I followed an object orientated approach for development. I instantiate a grid object. This grid object was assigned an array of hexagon objects, each with their own center coordinates.

Once I have my objects I assign a entity and there sentiment to each hexagon in the grid. To calculate what word should be assigned to what hexagon I use a distance formula. This ensures words with a higher sentiment positive or negative are further from the center. This calculation is trivial once each hexagon has its x,y coordinates for its location. I also colored the hexagons based on a normalized color spectrum to indicate the level of sentiment.

When all the required information is gathered and structured, I then draw the hexagons onto the canvas. One intertesing point is that each hexagon is drawn from two seperate hexagons. Using two hexagons gives the bar hexagon seen in the images.

The following code is how the hexagons are drawn onto the canvas.

function drawPloygon(hexagon) {
    ctx.lineWidth = hexagon.lineWidth;
    ctx.moveTo(hexagon.points[0].x, hexagon.points[0].y);
	for(var i = 1; i < hexagon.points.length; i++)
		var point = hexagon.points[i];
		ctx.lineTo(point.x, point.y);
    ctx.fillStyle = hexagon.colour;
    ctx.lineWidth = hexagon.lineWidth;

The follwoing is screenshot of the application of the finished design.


Access to data is key for any modern web application. For this reason I allowed the latest trending hashtags to be analyized and visualized through the application. I feed a significant number of tweets from twitter into Watson API’s and then visualized the results. This was all deployed on IBM Cloud using Cloud Foundry and its python runtime environment.