Is anyone else trying to use the Digistump Dashboard to display data from their Particle devices ?
I’m impressed by the integration, but after my initial success, I have not been able to get sparklines to actually plot on a widget. I can see the data is there, because the values show up when I cursor over them:
Is anyone here seeing anything similar ?
One possible item I should mention is that the left most data always seems to be not a number, “NaN” - I have nio idea if this is related, but I have no idea how to avoid that.
Most of the data I am reading is JSON encoded, and this fails on multiple browsers, but the snapshot above was taken from Firefox Version 47.0.
Erik from Digistump here - glad to hear someone is trying to use this with Particle devices as well.
Have you tried the other widgets - specifically the time series graphs? Those were tested the most, since we made them custom for this.
How about widgets like the text widget?
Just want to establish if it is data source related or widget specific before I dive into it (because I’m spread a bit thin, but really want this to be useful to all).
The code is also on github (github.com/digistump/OakDashboard) and can be tested locally with only a browser if you feel like diving deeper into it yourself.
FWIW - I do not think it is a data source problem (or at least not a straightforward one), because gauges work, and also I see the datapoints on the sparklines, just not the lines.
Well it would certainly help if the time series graphs actually showed up on the list-
They should now.
I’ve also implemented some changes that should prevent a datasource that has an initial datapoint that isn’t a number from breaking the widget (implemented for sparkline and time series graphs).
A few other little tweaks too that should make things smoother with displaying particle data.
Time series and single metric sparklines seem to be working now - but multi-metric sparklines are broken (and I can’t even see the data points when I hover over the widget now.)
Sparklines should work with multiple metrics now - and should not break with datasets that produce non-numeric data at the start, but I haven’t been able to reproduce that part so that is only theoretical
Is Digistump associated with Particle? I want to try the dashboard but I am cautious about where I log into using my particle account name and password.
Is there a place to see examples of a dashboard or a youtube type video that you can watch without supplying the username and password first?
Digistump partnered with Particle to bring the Particle Cloud to our product the Digistump Oak.
But no need to worry how we are connected because we don’t capture your login info in any way - you are logging into Particle’s servers directly from your browser, it never crosses our servers. Particle returns your Auth Token to this login request, also only to your browser, so that the dashboard can use it. Your auth token never passes through our servers either, which is why you have to login every time you use it.
The only thing that goes to our servers is when you hit save to server, we save the setup information for your dashboard - not any of the data or login information, just what widgets, where you placed them, and what settings you configured for them. We save that data to Amazon S3 to be exact, and don’t use it in anyway except to present the dashboard that was saved when the unique link is visited.
You can confirm any of this with traces, reading the source code, or looking at the project here: github.com/digistump/OakDashboard - it is open source of course.
Thanks for the reply I did a little more research after posting the question and had decide it was safe to login. My intent wasn’t to suggest that it wasn’t safe but only to double check. Sometimes a little caution can save a lot of problems, but to much caution can prevent a lot of fun.
However, I do have a major problem with these browser-based solutions, precisely because they are browser-based.
There is no history, you take a laptop/tablet and put it to sleep - the data that occurred while you are asleep is unavailable, and the discontinuities in the time-series & sparkline charts is confusing/misleading.
I understand that storing data and replaying it when dashboards (re)attach costs storage (i.e. $$), but without that capability these dashboards are doomed to remain toys.
I will state that I really like the integration with the particle cloud - specifically the parsing of JSON data from publish() events, that is a massive step forward. The optimal use model, IMHO, is to have sensors publishing data in JSON format, that can be consumed by webhooks as well as tools like this. Leverages the cloud infrastructure and scales from photons, to Bluz, to digistump oaks, to electrons.
I would like to be able to find a way to embed an auth token in the shortcut, so that the manual step of logging in could be avoided (maybe that’s already there but I haven’t mastered it yet.)
Here’s an example of the browser-based problem I refer to. Look at the timestamps on the time series charts, and the discontinuities in those and the sparklines, misleading or just plain wrong depending on how important your data is.
@digistump could add cookies or local storage to it to store the accesstoken locally. I've been using that for about a year on my 'interface' and it's working lovely. Not sure how much of a security issue that is, but it's only on your local machine, so should be reasonable(?)
As for the graphs, perhaps it's best to show no line for when there's no data. That'd leave a gap, but it's then clear that data was unknown at that time, rather than giving guesses.
@carbuthn - no manual but it is based on freeboard.io so anything answers you can find for that apply. To answer: 1. Not sure why you’re saying they are strings, decimal numbers are working in my testing. 2. When you fill in the data box on the widget displaying the data you can use javascript in that box - there is a link to bring up a JS Editor box too, but you can just type it right in - for example something like Math.round() should work around your datasource name.
@AndyW - Glad you like the integration. No plans to add data storage to it, but I fully plan to support pulling from data storage when Particle adds that as part of the Particle Cloud.
@AndyW and @Moors7 - I’ve added the ability to set an auth token and skip the login - just add ?auth=XXXX to the end of the URL where XXXX is a valid token for your Particle account. I’ve made it show this option when you save the current dashboard to the server. I don’t plan to extend that to cookies, or any other sort of storage as the user needs to be very aware that with that auth token someone can do anything to their devices. If Particle adds read only auth token in the future than having a link to a “public” copy of a dashboard will be possible to add.
Thanks,
The login with the token will be a big help.
I have never used java script before, guess that is one more thing to come up to speed on. After a little bit of looking I did find that I could add “.toFixed(1);” to the end of the data source and limit the number of digits after the decimal point.
