Timeline for Reducing number of data points in excel while keeping the curve shape
Current License: CC BY-SA 3.0
41 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 3, 2018 at 20:49 | answer | added | AdamO | timeline score: 1 | |
| Jul 2, 2018 at 1:03 | comment | added | Glen_b | It's a couple of years ago (so I don't recall exact details), but I'd say I just used some drawing package that allows you to manipulate 2D linear splines (many do) to lay down a set of points close to the observed series and then pushed the points around until they looked as near to the line as I could get without spending too much time on it (i.e. quite literally judged it by eye). The point was simply to illustrate what such a broken-line fit would achieve, not to actually estimate one (though packages to do this exist), so that the OP could indicate whether that would be sufficient. | |
| Jul 1, 2018 at 18:13 | review | Close votes | |||
| Jul 3, 2018 at 20:38 | |||||
| Jul 1, 2018 at 18:00 | comment | added | Jim | @Glen_b may I ask how you achieved that "by-eye broken-line fit" with OP's data? | |
| Jul 1, 2018 at 17:35 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jun 1, 2018 at 3:54 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Apr 30, 2018 at 11:38 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Mar 30, 2018 at 6:54 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Feb 25, 2018 at 2:11 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jan 24, 2018 at 2:34 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Dec 24, 2017 at 5:52 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Nov 2, 2017 at 1:22 | history | edited | Ferdi |
edited tags
|
|
| Nov 2, 2017 at 1:10 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Sep 28, 2017 at 8:53 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Aug 20, 2017 at 17:33 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jul 19, 2017 at 0:50 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jun 18, 2017 at 13:32 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| May 8, 2017 at 16:09 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Apr 5, 2017 at 20:44 | comment | added | whuber♦ | The solutions to the (remarkably similar) question at stats.stackexchange.com/questions/35220 might be helpful. I suspect my second answer at stats.stackexchange.com/a/35268/919, which is fully automatic, might work nicely here, too. | |
| Apr 5, 2017 at 18:26 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Feb 22, 2017 at 23:12 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| S Sep 29, 2016 at 13:43 | history | bounty ended | CommunityBot | ||
| S Sep 29, 2016 at 13:43 | history | notice removed | CommunityBot | ||
| Sep 28, 2016 at 1:04 | answer | added | hyiltiz | timeline score: 0 | |
| Sep 25, 2016 at 21:54 | comment | added | Carl | It is not clear what you need. Are the local variations noise or data? If they are useful data, then the procedure is different than if they are just noise. | |
| Sep 24, 2016 at 9:54 | comment | added | tomka | I just started an answer to your Q but realized I have insufficient information. I need two pieces of information. 1) What is the variable on the y-axis? Is this a count or other distributional information (and hence the plot is some sort of histogram with kernel density estimator) or is this a second variable (and hence the plot is a scatterplot). 2) In case it is a scatterplot it looks a lot like time series, in which case x would be time and y the level of the outcome. Is this time-series data? | |
| Sep 22, 2016 at 4:26 | comment | added | GeoMatt22 | When you say "in Excel", that tends to limit the answers you will get. In terms of built-in functionality, it is easy to do linear/semilog/log regression, and I'm not sure what else. If VBA is on the table, then there are more possibilities. (I do not use VBA myself, but I believe it runs natively in Excel, and for example this appears to be a VBA version of the standard curve simplification algorithm I mentioned above.) | |
| Sep 21, 2016 at 14:39 | history | tweeted | twitter.com/StackStats/status/778604447337439232 | ||
| Sep 21, 2016 at 10:24 | comment | added | Gumeo | What would you use to measure the quality of the down-sampled curve? I think that if you can formulate that more specifically other than visual inspection, then we can help you better. | |
| S Sep 21, 2016 at 10:02 | history | bounty started | Glen_b | ||
| S Sep 21, 2016 at 10:02 | history | notice added | Glen_b | Draw attention | |
| Sep 20, 2016 at 0:51 | history | edited | Kegan | CC BY-SA 3.0 |
added 544 characters in body
|
| Sep 20, 2016 at 0:37 | comment | added | Kegan | I updated the question, so I hope that helps with conveying what I need to do. @Glen_b The little wiggles (which pretty much look like noise) I don't care about keeping, what I want to preserve is the bigger irregularities. Here's the points of data I gathered by hand before looking for a better method. | |
| Sep 20, 2016 at 0:27 | history | edited | Kegan | CC BY-SA 3.0 |
added 1263 characters in body
|
| Sep 18, 2016 at 5:40 | comment | added | Glen_b | Your question isn't yet clear enough to sure what you need but you might do okay with a natural cubic spline on a log-scale. If you want to be able to do linear interpolation, an ordinary linear spline might do well enough. Here's a by-eye broken-line fit with only 13 points -- if that's the sort of thing you need there are a number of ways you might achieve something like it. | |
| Sep 18, 2016 at 4:46 | comment | added | Glen_b | When you say "keep the shape of the curve", the displayed curve shows a lot of wiggles. You're going to have to be considerably more precise about what you want to keep and what you don't. | |
| Sep 17, 2016 at 22:21 | comment | added | Pere | If you want to smooth the curve while retaining its shape, you can try using the mobile average (en.wikipedia.org/wiki/Moving_average) instead of the original data. It should be easy to be done in Excel. | |
| Sep 17, 2016 at 21:57 | comment | added | GeoMatt22 | I agree with @MatthewDrury, and for a simple Excel solution it looks like a power law might do OK. If you have to sub-sample the original data, a standard approach is curve simplification, but not easy to do in Excel (presumable would require VBA?). | |
| Sep 17, 2016 at 21:40 | comment | added | Matthew Drury | Do the data points you keep have to come from the origional data set? If not, a good bet would be to fit a curve to your data, then use that fit curve to generate new data points that exactly capture the trend. | |
| Sep 17, 2016 at 21:23 | review | First posts | |||
| Sep 17, 2016 at 21:34 | |||||
| Sep 17, 2016 at 21:20 | history | asked | Kegan | CC BY-SA 3.0 |
