TauCharts: Data-focused JavaScript charting library | Hacker News Reader

TauCharts: Data-focused JavaScript charting library

blog.taucharts.com

104 pointsinfinite8s39 comments

At the http://taucharts.com it says "Free open source D3.js-based library" but in the tutorial at http://api.taucharts.com/tutorials/5min.html it says "free for non-commercial use". At GitHub it is sporting Apache License, Version 2.0 which - if I am not mistaken - allows commercial use. Could you please clear that up for me? Am I allowed to use it in commercial products (like a website made for one of my clients) or not? Thanks.

It is free for commercial use, the site will be fixed, it is a mistake in the 5min page.

Great, I see that you already fixed it, thank you - I can add "provider solving problems quickly" to my personal list of reasons to use the app.

I still find charting tools that do not allow interaction with the a group of charts to enable further filtering very lacking. The gold standard I evaluate any charting library against is http://dc-js.github.io/dc.js/

My reasoning is that most dashboards should be designed to enable data to tell a story and enable users to interact and create their own story.

As a taucharts developer, I would give you a "best comment" award :) Actually taucharts is moving to this direction. We want to simplify API for dashboards and composite charts to allow users create "visual stories". Stay tuned :)

I think the reason dc.js works so well is that it is tightly coupled to crossfilter (http://square.github.io/crossfilter/). This manages the multi-dimensional views that dc.js leverages to render charts.

When developing, the chart bit is usually the easiest bit to implement. Getting the data set into the views you need is where you spend a lot of time.

Good luck with the chart engine!

That or NVD3 (http://nvd3.org/) They didn't take it as far as DC.js but it does allow at least some sort of filtering like options.

(Thanks for the link. DC looks awesome.)

This is the feature that most users of QlikView, PowerView and Tableau are buying from an end user experience point of view.

There are essentially two types of dashboard experiences:

1. Static: a collage of individual visual elements presented as on a single page;

2. Interactive: the same as above but with the capability to do filtering.

Please don't use the combination of red and green as early option for category colors. Many people have color vision problems and that combinations is very problematic for them.

Thanks for notice, we will think about that. Actually chart API allows to configure color brewer and I hope we will have several themes soon.

This is a rather minor nitpick, but I think the documentation and TauCharts website as a whole could use some editing by a native English speaker.

That said, the content and goals for the project really are quite cool! I'm looking forward to where TauCharts can go in the future.

If your goal is to express ideas and get someone to contribute/download/convert, I don't think that's a minor nitpick at all. It was pretty painful to read certain parts.

As well documentation for taucharts is open sourced: https://github.com/TargetProcess/taucharts-docs. You can help a project.. just send pull requests with corrected grammar.

Really sad to hear that :/ but you right. Will improve documentation for sure

What makes it better than Vega? I like how straight forward it is, much like highcharts. I'll explore it more tomorrow.

Actually I personally like the concept of Vega but in my humble opinion the protocol is too academic, implementation goes too slowly so I doubt that Vega suits for real world application.

It seems interesting for learning purpose.

So, if I get this right, being 'data-friendly' sums up to using data in the form [ {val1: 1, val2: 2}, {val1: 2, val2: 3}] and then mapping x: 'val1', y: 'val2' instead of something like {x: [1,2], y: [2,3]}. The former assumes that the input data comes in a csv or a similar format. This makes charting computed data (à la MATLAB or numpy/matplotlib) rather awkward.

You've got it right. Also I have an idea to make an API that joins data processing and plotting like tauCharts.data(#).aggregate(#).plot(#)

I like this. A 20 second glance at the JavaScript in the first example was all it took for me understand how this works.

What I want with such charting solutions is something to export or copy and paste data from excel and put it into a chart. Just that. No need to manually edit a javascript file or html code beyond customising the appearance once the data is in there. Is this available somewhere?

https://plot.ly/

Some of our biologists use plot.ly. They seem to like it.

Those that know it still are using ggplot2 and R. This grammar of graphics plotting library really seems to gold standard of plotting.

D3 has an importer for tabular data, and looks like Taucharts uses D3 under the hood.

That said, it would be cool if there was some kind of composer interface that would allow you to point-at/paste/import data, then save the visualization as a single page HTML snippet - like a small Tiddlywiki for data.

- https://github.com/mbostock/d3/wiki/CSV - http://tiddlywiki.com/

We are working on such a solution http://vizydrop.com/

Actually it uses Taucharts under the hood.

Cool. Look forward to hearing more.

https://tuvalabs.com

Easy to import the data and allows you to visualize the information with various different chart types.

My first impression could not be better - it seems to be simple to use and it looks good. BTW, when you click on the Facebook button at the bottom of the page, the sharing box is under that disqus box - you might want to fiddle with z-index or something.

Thanks for positive feedback. FB button is already fixed

I wonder how this compares to VarianceCharts - https://variancecharts.com/ (besides the license). They both claim to get inspiration from the grammar of graphics.

As far as I understand from quick review the projects are quite similar. But variancecharts don't have facets and not free for commercial use. Difference in API doesn't matter. Taucharts can be wrapped to webcomponent and described on the page using tag as well.

What's most astonishing about all this is that the AMC Rebel gets 0mpg!

I'm happy you noticed that :) Obviously this is a corrupted row in a data so you can try "exclude" feature of tooltip plugin to remove this record from the chart in one click

Any plans for spider chart? Geo charts?

Basic Geo charts are already there, but not documented so far.

In future we are going to support all type of visualizations from Grammar of Graphic, so it includes spider chart as well. In fact it is Polar Parallel chart and it will be possible to create it when support of polar coordinates will be added.

Given that you called them facets, I'm going to assume you're familiar with ggplot2. In retrospect, I thinking adding non-Cartesian coordinate systems was a mistake. They are a huge amount of work and only make a handful of charts easier (and most of which, like the radar chart, are not that useful).

how much data you can visualize with TauCharts? And on the first glance couldn't see how many different types of chart you can make.

Friendly speaking, we didn't make a performance or optimization research yet. I can only share an experience we've got from using taucharts in our main product - "Targetprocess". So we setup a limit to 1000 rows (+/-50%) which seems to be acceptable performance / usefulness balance to draw a chart within a 0.5 seconds. Also it depends on chart type. Drawing huge facet with more than 100 cells in it can be painful while plain scatter plot on the same data works fine.

Currently taucharts supports scatter, bar (vertical / horizontal), stacked bar (vertical / horizontal), line. Also in beta we have geomap charts with customized map and parallel coordinates. There is a lack of documentation.. but we are working on this problem.

Reply on news.ycombinator.com