That looks fine. How are you limiting the results when selecting the different scales, are you just assuming there will always be a maximum number of readings a day and adjusting the limit?

Yeah maximum plus 20 more to give few extra at the start.

1hour - 80
2 hours - 140
6 hours - 380
12 hours - 740
1 day - 1440
1.5 days - 2180
2 days - 2900
7 days - 10100

It defaults to 2900

To me graphs are about being able to see trends and anything less than 1 day doesn't really tell you much so I might get rid of them when I have a decent amount of data.

• Good

1. I'm glad

Soon you will be able to tell us all that it has been the hottest day in the history of hot!

FWIW: http://caer.me:2988/server.js

That's the source to the Node app that's serving the monitoring page.

Also, in the time between writing the above paragraph and starting to write this one, I rewrote a big chunk of both the monitoring, server and client-side code. The monitoring code now stores counts etc. in the database (previously it appended to a log file), while the server code now sends the data in a nicer format. The client code has been updated to take that into account, and is now simpler as a result.

Oh and Redis is cool, you should try it. Much more lightweight and straightforward than *SQL.

edit:

oops, just noticed my MIME type thing is a bit stupid.

edit 2:

How come you're not requesting the graph data via AJAX? Seems a simpler way to do it, imo, and less for your server to do.

AJAX...because I don't know how to. I've never knowingly used any AJAX code in anything I've made. Basically all I tend to do when I make something online is tack together the bits of things I already know and I've already made else where and then learn a little bit more to get whatever I want working sufficiently :$

It's really simple. First you make a PHP thing to serve the data:

<?php
// this is get_data.php
$time = $_GET['time'];
$dataset1 = magic_function_to_fetch_outdoors_data_from_db($time);
$dataset2 = magic_function_to_fetch_inoors_data_from_db($time);

$data = array(
  "dataset1" => $dataset1,
  "dataset2" => $dataset2
);

echo json_encode($data);
?>

Then you modify your HTML links like so:

<ul id="period">
  <li><a href="get_data.php?time=1h">1 hour</a></li>
  <li><a href="get_data.php?time=2h">2 hour</a></li>
  <li><a href="get_data.php?time=6h">6 hour</a></li>
</ul>

Then you have some JS that looks something like this:

var $chart = $("#placeholder"),
    options = {
        lines: { show: true },
        points: { show: false }
        // etc.
    };

function plotChart(data) {
    $.plot($chart, [
        { data: data.dataset2, color: "#DE1B55", label: "Indoors" },
        { data: data.dataset1, color: "#5F2BCF", label: "Outdoors" }
    ], options);
}

$("#period").on("click", "a", function () {
    var $link = $(this);

    $link.addClass("working");

    $.ajax({
        url: this.href,
        type: "GET",
        dataType: "json"
    }).done(function (data) {
        plotChart(data);
    }).fail(function () {
        alert("O noes!");
    }).always(function () {
        $link.removeClass("working");
    });

    return false; // stops the browser from trying to follow the link
});

And that's basically it.

Not really directly relevant but I saw this today, seems pretty coool: https://plot.ly/plot

Might you not speed it up with fewer samples? Every minute seems excessive (maybe you needed that for developing/debugging).

That choked my dual core 2.5g.

There isn't really any developing/debugging as such. I've been playing with temperature sensors with my RPis and I wanted to put one outdoors and I decided it would be nice to look at the data in a graph. Every minutes is definitely overkill but it's very quick when I access the data locally, I think it would be slow even if the data was held on the hosting server, especially when it gets to weeks and months worth of data.

At that rate, 24 hours' worth is only 60k or so – 1440 data points per series, which is not a huge amount, and like I said earlier, even my phone can handle that with no noticeable slowing.

No drop shadow! :O

A MySQL question.

Due to there being missing parts on the chart for the shorter readings because of the missing entries I've changed how the data is pulled from the database by calculating the current time minus the length of time required then only fetching the rows that are greater than this calculated time. It works nicely and gives a proper day, week length etc.

My question is should I be indexing my "unixtime" column to make it easier on the DB? 

New code is...

$currtime = time();

$newlimit = $currtime - $limit;

SELECT * FROM outside WHERE unixtime > $newlimit ORDER BY id DESC

I suck at using the new editor.

Yes, an index would be useful there.

The editor sucks at being used, more like.

Is it a cache issue that is causing no grey box to appear when I want to add code in the new editor? Only "code:" appears I end up having to go in to the HTML and put things between the pre tags but there's divs all over the place.

Could be a cache issue, it's showing here on both "Teh Forum" and "Yellow" themes.

Another slightly irrelevant cool thing: http://ankane.github.io/chartkick/