Need Help with CDF
arilwan opened this issue · comments
HI @AllenDowney,
I have been trying to use CDF plot for a multiclass dataframe, but all efforts failed. Here's how my dataframe looks like:
data.shape
(4672, 6)
data.head()
+---+---------+----------+----------+-------+--------------+------------+
| | trip_id | duration | distance | speed | acceleration | travelmode |
+---+---------+----------+----------+-------+--------------+------------+
| 0 | 303633 | 2657 | 5624.06 | 2.12 | 0.00080 | bus |
| 1 | 303637 | 1185 | 1274.04 | 1.08 | 0.00091 | foot |
| 2 | 303638 | 1185 | 4464.56 | 3.77 | 0.00318 | car |
| 3 | 303642 | 3350 | 7715.78 | 2.30 | 0.00069 | bus |
| 4 | 303657 | 704 | 1155.26 | 1.64 | 0.00233 | car |
+---+---------+----------+----------+-------+--------------+------------+
Now I want 4 CDF plots each for duration, distance, speed, and acceleration for the 5-class travelmode in the dataset [bike, bus, car, foot, metro]. I have been trying this for two days, but mess-up.
My objective is to plot (for each of duration, speed, distance, acc.) a distribution by mode, something like the following:
Any help please?
If you can send me starter code, I might be able to work up an example.
But I would start with
```
for name, group in data.groupby('travelmode'):
Cdf.from_seq(group.duration).plot()
```
On Mon, Nov 4, 2019, at 5:47 PM, Akilu Rilwan Muhammad wrote:
HI @AllenDowney <https://github.com/AllenDowney>,
I have been trying to use CDF plot for a multiclass dataframe, but all efforts failed. Here's how my dataframe looks like:
`data.shape
(4672, 6)
data.head()
+---+---------+----------+----------+-------+--------------+------------+
| | trip_id | duration | distance | speed | acceleration | travelmode |
+---+---------+----------+----------+-------+--------------+------------+
| 0 | 303633 | 2657 | 5624.06 | 2.12 | 0.00080 | bus |
| 1 | 303637 | 1185 | 1274.04 | 1.08 | 0.00091 | foot |
| 2 | 303638 | 1185 | 4464.56 | 3.77 | 0.00318 | car |
| 3 | 303642 | 3350 | 7715.78 | 2.30 | 0.00069 | bus |
| 4 | 303657 | 704 | 1155.26 | 1.64 | 0.00233 | car |
+---+---------+----------+----------+-------+--------------+------------+
`
… Now I want 4 CDF plots each for duration, distance, speed, and acceleration for the 5-class `travelmode` in the dataset `[bike, bus, car, foot, metro]`. I have been trying this for two days, but mess-up.
My objective is to plot (for each of duration, speed, distance, acc.) a distribution by mode, something like the following: Screenshot 2019-11-04 at 22 43 34 <https://user-images.githubusercontent.com/42409395/68164543-fb4d6580-ff54-11e9-800c-ab13a3570841.png>
Any help please?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AAOLP3O2DZ7B4T6IMELBEV3QSCRBXA5CNFSM4JI2VND2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HWXLFMA>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOLP3JVMZUYIQL5ZOWC3TTQSCRBXANCNFSM4JI2VNDQ>.
Does that mean you figured it out? If so, well done!
…On Tue, Nov 5, 2019, at 6:50 PM, Akilu Rilwan Muhammad wrote:
Closed #19 <#19>.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AAOLP3NP3QCW3DNBPV7FVSDQSIBGFA5CNFSM4JI2VND2YY3PNVWWK3TUL52HS4DFWZEXG43VMVCXMZLOORHG65DJMZUWGYLUNFXW5KTDN5WW2ZLOORPWSZGOUVHTMZI#event-2773431909>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOLP3LWV74F7FLQDEKACSDQSIBGFANCNFSM4JI2VNDQ>.
Yes, many thanks.
I'm very sorry to reopen this issue, I'm in need of help with the plot, I couldn't manage to add a legend to the plot to show which plot is for which mode?
for name, group in data1.groupby('travelmode'):
Cdf.from_seq(group.speed).plot()
title, x, y = 'Speed by mode','speed (km/h)', 'CDF'
decorate_cdf(title,x,y)
Something similar to the first plot above.
When you call `plot` on a `Cdf`, you can calling plot on a Pandas series. It takes the same keyword arguments, including `label`.
…On Mon, Dec 2, 2019, at 5:48 AM, Akilu Rilwan Muhammad wrote:
Reopened #19 <#19>.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AAOLP3NSACGCQ75E5K5PE2TQWTRYTA5CNFSM4JI2VND2YY3PNVWWK3TUL52HS4DFWZEXG43VMVCXMZLOORHG65DJMZUWGYLUNFXW5KTDN5WW2ZLOORPWSZGOVGLANKY#event-2845181611>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOLP3JWKOB2BHG54USFI6TQWTRYTANCNFSM4JI2VNDQ>.
Oh yes, adding
plt.legend(groups.groups.keys())
to my decorate_cdf function did the trick.