# Include graphs of previous trend on the dashboard
**URL:** https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398
**Category:** Feature
**Created:** [5월 11, 2018, 3:13오후 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398 "2018-05-11T15:13:03Z")
**Posts on this page:** 11
**Page:** 1
### Author:  [@mcwumbly](https://meta.discourse.org/u/mcwumbly)
#### Post date: [5월 11, 2018, 3:13오후 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/1 "2018-05-11T15:13:03Z")
> [@Help us build the new 2.0 Dashboard](https://meta.discourse.org/t/help-us-build-the-new-2-0-dashboard/86468/78):
>
> - Trends are now calculated as compared to previous period. So when looking at trend on a month it compares to last month. When looking at trend on a week it compares to last week.
This is nice, thanks. It’d be even better if the previous period were charted as well, like what Google Analytics does, for example, (dotted line for previous period vs. solid line for current period):
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [5월 14, 2018, 5:49오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/2 "2018-05-14T05:49:22Z")
I think this is a great suggestions for a future version of the dashboard, in fact it is so good I am slotting this for 2.1
---
### Author:  [@erlend\_sh](https://meta.discourse.org/u/erlend_sh)
#### Post date: [9월 5, 2018, 8:32오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/3 "2018-09-05T08:32:26Z")
This is done
---
### Author:  [@codinghorror](https://meta.discourse.org/u/codinghorror)
#### Post date: [9월 5, 2018, 8:37오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/4 "2018-09-05T08:37:56Z")
Are you sure this is useful @mcwumbly looking at the data?
---
### Author:  [@mcwumbly](https://meta.discourse.org/u/mcwumbly)
#### Post date: [9월 5, 2018, 8:51오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/5 "2018-09-05T08:51:50Z")
No, it’s not obviously useful at all at the moment. I can’t tell if that’s because the idea is bogus or something is off in the implementation yet, though.
I thought it was just me at first, but looking at the screenshot above, it corroborates that for whatever reason, the data dotted line is only visible for two graphs: Signups and New Contributors. What about the others?
Another thing that jumps out at me from looking at these is the likely value in standardizing on periods that are multiples of 7 days. If the above graphs were for 28 day periods instead of 30-31 day periods, I bet the natural peaks and values of the weekdays/weekends would line up better and the comparison would make more sense.
---
### Author:  [@codinghorror](https://meta.discourse.org/u/codinghorror)
#### Post date: [9월 5, 2018, 8:54오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/6 "2018-09-05T08:54:55Z")
Yeah the lack of “lining up” makes it particularly useless, so we should fix that @j.jaffeux
---
### Author:  [@merefield](https://meta.discourse.org/u/merefield)
#### Post date: [9월 5, 2018, 9:14오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/7 "2018-09-05T09:14:30Z")
On a related note I originally proposed a trend line might be added which used two or more moving averages across the graph to show general trend and would cut through the noise of the dips and peaks. Obviously the monthly or weekly aggregation goes some way towards that but wondered if there was still appetite for a ‘straight line fit’? e.g.:
---
### Author:  [@mcwumbly](https://meta.discourse.org/u/mcwumbly)
#### Post date: [9월 5, 2018, 9:29오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/8 "2018-09-05T09:29:29Z")
I sketched something similar out in the past here:
> [@Encouraging User Participation - Some Ideas, More welcome](https://meta.discourse.org/t/encouraging-user-participation-some-ideas-more-welcome/42511/75):
>
> Played around a little with smoothing today, in case anyone wants to explore further: var smooth = function(data, windowSize) { var smoothed = []; var movingWindow = []; for(var i = 0; i \< windowSize; i++) { movingWindow.push(data[i]); } for(var i = 0; i \< data.length; i++) { movingWindow.push(data[i]); movingWindow.shift(); smoothed.push(movingWindow.reduce((a, v) =\> a + v, 0) / windowSize); } return smo…
---
### Author:  [@codinghorror](https://meta.discourse.org/u/codinghorror)
#### Post date: [9월 5, 2018, 10:02오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/9 "2018-09-05T10:02:09Z")
Just as a reminder my original use case was putting two graphs with independent Y-axes together, namely posts and replies. Because there is a clear relationship between the two – you could theoretically have one topic with 10,000 replies in a week, or you could have 10,000 topics with 1 reply each 😉
---
### Author:  [@j.jaffeux](https://meta.discourse.org/u/j.jaffeux)
#### Post date: [9월 5, 2018, 9:36오후 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/10 "2018-09-05T21:36:54Z")
I deactivated this for now. I have multiple ideas to make it work, but this is not the right time. Also I see very different things in the topic, maybe we should discuss this more post release.
---
### Author:  [@loginerror](https://meta.discourse.org/u/loginerror)
#### Post date: [4월 17, 2020, 9:39오전 UTC](https://meta.discourse.org/t/include-graphs-of-previous-trend-on-the-dashboard/87398/11 "2020-04-17T09:39:43Z")
단순한 아이디어 하나를 떠올렸습니다. 그래프를 잡고 가로로 스크롤하여 더 많은 과거 결과를 표시할 수 있다면, 단기 통계를 빠르게 파악하는 재미있고 동적인 방식이 될 것 같습니다 🙂