search

monroews / CEE4530

MIT License
4 stars 11 forks source link

Curve fitting #128

Closed jacqueline-wong closed 5 years ago

jacqueline-wong commented 5 years ago

I used Excel to find a curve that fit nicely with the first 6 points and plotted another line to plot the last few (which all have values of 0). This was a polynomial (to the power of x**4), which only works empirically to fit this data. Since there is not much explanation for why this is the case, should I... 1) Leave it as it is and just explain that it was the curve that fit best 2) Not put a curve at all because I do not know how to explain the model? Thank you!

Screenshot 2019-05-10 at 5 01 39 PM
jacqueline-wong commented 5 years ago

Similarly for here (which is volume of water treated for the first 6 experiments shown above), should we be using a linear fit or a logarithmic fit? The logarithmic fit looks nicer but again, I cannot justify any of the results.

Screenshot 2019-05-10 at 5 27 40 PM
monroews commented 5 years ago

It is fine to explain what you said in your question. The curve fit doesn't make sense, and it shows the data trend nicely. This is great data and you can try to develop an explanation!

On Fri, 10 May 2019, 5:28 p.m. lw583, notifications@github.com wrote:

Similarly for here (which is volume of water treated for the first 6 experiments shown above), should we be using a linear fit or a logarithmic fit? The logarithmic fit looks nicer but again, I cannot justify any of the results. [image: Screenshot 2019-05-10 at 5 27 40 PM] https://user-images.githubusercontent.com/46941988/57557647-ee0a4200-7348-11e9-946a-f9e0fd6d01c8.png

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/monroews/CEE4530/issues/128#issuecomment-491435614, or mute the thread https://github.com/notifications/unsubscribe-auth/AGZX725PEFU5IJZXZNCDLATPUXSITANCNFSM4HMGGZVQ .

monroews commented 5 years ago

Consider using time/hydraulic residence time on the y axis of the first graph

On Fri, 10 May 2019, 7:42 p.m. Monroe Weber-Shirk, monroews@gmail.com wrote:

It is fine to explain what you said in your question. The curve fit doesn't make sense, and it shows the data trend nicely. This is great data and you can try to develop an explanation!

On Fri, 10 May 2019, 5:28 p.m. lw583, notifications@github.com wrote:

Similarly for here (which is volume of water treated for the first 6 experiments shown above), should we be using a linear fit or a logarithmic fit? The logarithmic fit looks nicer but again, I cannot justify any of the results. [image: Screenshot 2019-05-10 at 5 27 40 PM] https://user-images.githubusercontent.com/46941988/57557647-ee0a4200-7348-11e9-946a-f9e0fd6d01c8.png

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/monroews/CEE4530/issues/128#issuecomment-491435614, or mute the thread https://github.com/notifications/unsubscribe-auth/AGZX725PEFU5IJZXZNCDLATPUXSITANCNFSM4HMGGZVQ .

jacqueline-wong commented 5 years ago

I plotted it with the normalized time but decided to keep it just as time because I also have a graph of volume treated, which looks almost exactly the same as the one with normalized time! I wanted to show what it looked like before and after normalization in my analysis. Hope that is a good enough reason. Thanks!