I have written code for customizing my x ticks, snippet of the same is below
arr_label = ['sum_msg_len','log_count','info_hit','debug_hit','error_hit']
for label in arr_label :
fig = plt.figure(figsize=(15,6))
axes = fig.add_axes([1,1,1,1])
axes.xaxis.set_major_locator(plt.LinearLocator(30))
axes.tick_params(axis ='x',labelsize=6)
axes.plot(df.index,df[label],'g',label =label)
axes.legend()
fig.autofmt_xdate()
fig.savefig('images_indv/'+app_index+"_"+label+".png",bbox_inches='tight')
#fig.close()
fig.clf()
my requirement is that is have timestamps spaced by minute and i want to plot timestamp vs ('sum_msg_len'/'log_count'/'info_hit'/'debug_hit'/'error_hit') one by one,
but problem is X ticks, i want some specified no of ticks to appear within the range of the data which i am plotting.
Earlier when i was not specifing any Locator then all the timestamps got overlapped and one cannot read the timestamps properly. So when i try to use a locator, it labels the x-axis with out any relation to the plotted value.
Like if i use LinearLocator(30) it just plots the first 00 to 29 mins in the graph,and if i use LinearLocator(50) it just plots the first 00 to 49 mins in the graph with no change to the y axis values. Plots of both I am putting below. I also tried with different locators Like MultipleLocator and MaxNlocator, but issue sustains
In short, I just want the graph plotted for 21July 00:00:00 to 22 July 00:00:00 which will be 1440 entries but the i want to see around 30-40 intermediate entries mentioned on the plot.
Related
How can I plot the graph
Getting the data from those 3 sources
Using only first letter and last digits of the first column to put it in the X-axis as in the Excel graph above
How can I only show first column data by 20 digits difference ? aa010 aa030 aa050 ... etc
I have three different data from a source. Each one of them has 2 columns. Some of those 3 sources' first columns named the same but each one of them has different data corresponding to it in the second column.
I need to use python to plot those 3 data at one graph.
X-axis should be the combination of the first column of three data from the sources. - The data is in format of: aa001 - (up to sometimes aa400); ab001 - (up to sometimes ab400).
So, the X-axis should start with a aa001 and end with ab400. Since it would just overfill the x-axis and would make it impossible to look at it in a normal size, I want to just show aa020, aa040 ..... (using the number in the string, only show it after aa0(+20) or ab0(+20))
Y-axis should be just numbers from 0-10000 (may want to change if at least one of the data has max more than 10000.
I will add the sample graph I created using excel.
My sample data would be (Note: Data is not sorted by any column and I would prefer to sort it as stated above: aa001 ...... ab400):
Data1
Name Number
aa001 123
aa032 4211
ab400 1241
ab331 33
Data2
Name Number
aa002 1213
aa032 41
ab378 4231
ab331 63
aa163 999
Data3
Name Number
aa209 9876
ab132 5432
ab378 4124
aa031 754
aa378 44
ab344 1346
aa222 73
aa163 414
ab331 61
I searched up Matplotlib, found a sample example where it plots as I want (with dots for each x-y point) but does not apply to my question.
This is the similar code I found:
x = range(100)
y = range(100,200)
fig = plt.figure()
ax1 = fig.add_subplot(111)
ax1.scatter(x[:4], y[:4], s=10, c='b', marker="s", label='first')
ax1.scatter(x[40:],y[40:], s=10, c='r', marker="o", label='second')
plt.legend(loc='upper left');
plt.show()
Sample graph (instead of aa for X-axis-> bc; ab -> mc)
I expect to see a graph as follows, but skipping every 20 in the X-axis. (I want the first graph dotted (symbolled) as the second graph but second graph to use X-axis as the first graph, but with skipping 20 in the name
First Graph ->- I want to use X-axis like this but without each data (only by 20 difference)
Second graph ->- I want to use symbols instead of lines like in this one
Please, let me know if I need to provide any other information or clarify/correct myself. Any help is appreciated!
The answer is as following but the following code has still some errors. The final answer will be posted after receiving complete answer at The answer will be in the following link:
Using sorted file to plot X-axis with corresponding Y-values from the original file
from matplotlib import pyplot as plt
import numpy as np
import csv
csv_file = []
with open('hostnum.csv', 'r') as f:
csvreader = csv.reader(f)
for line in csvreader:
csv_file.append(line)
us_csv_file = []
with open('unsorted.csv', 'r') as f:
csvreader = csv.reader(f)
for line in csvreader:
us_csv_file.append(line)
us_csv_file.sort(key=lambda x: csv_list.index(x[1]))
plt.plot([int(item[1]) for item in csvfile], 'o-')
plt.xticks(np.arange(len(csvfile)), [item[0] for item in csvfile])
plt.show()
I have a pandas dataframe of about 3 years with the resolution of 6 seconds and I want to group the data into 24-hour bins and plot each day using matplotlib in a loop.
This is my dataframe's head:
timestamp consumption
0 2012-11-11 12:00:03 468
1 2012-11-11 12:00:09 476
2 2012-11-11 12:00:16 463
3 2012-11-11 12:00:22 449
4 2012-11-11 12:00:28 449
It includes the power consumption of a house from 2012 till 2015. After the pre-processing, the dataframe starts at about 12 pm of the first day. I need to plot all of the dataframe in 24-hour intervals and each plot must represent for a single day that starts from about 12 pm and ends at about 12 pm of the next day
So, I need about 1500 plots that show the power consumption of each day starting from 12 pm, for about 1500 days of my dataframe.
Thanks in advance.
Update: The reason I want to plot 1500 days separately, is I want to check each night's power consumption and label the occupant's sleep pattern. And I considered each day from 12 pm to 12 pm to have a complete sleep cycle in one plot. And after preparing the labels I'll be able to use them as train and test data for classification
Consider this not only an answer but also a suggestion. First, convert the column 'timestamp' into the index (DatetimeIndex)
df.set_index(df['timestamp'], inplace=True, drop=True)
Then, get all the unique days that happen in your DataFrame
unique_days = list(set(df.index.to_period('D').strftime('%Y-%m-%d')))
We then squeeze the DataFrame into a Series
del df['timestamp']
df = df.squeeze()
Now, just plot unique days in your series in separate subplots.
import matplotlib.pyplot as plt
unique_days = list(set(df.index.to_period('D').strftime('%Y-%m-%d')))
fig, axes = plt.subplots(nrows=len(unique_days), ncols=1)
row = 0
for day in unique_days:
df[day].plot(ax=axes[row], figsize=(50,10))
row += 1
plt.show()
Now, it's time for you to play around with the parameters of plots so that you can customize them to your needs.
This is kind of a strange request. If we knew what your end objective is, it might be easier to understand, but I'm going to assume you want to plot and then save figures for each of the days.
df['day'] = (df['timestamp'] + pd.Timedelta('12h')).dt.date
for day in df['day'].unique():
mask = (df['day'] == day)
#<the code for the plot that you want>
plt.plot(x=df[mask]['timestamp'].dt.time,y=df[mask]['consumption'])
plt.savefig('filename'+str(day)+'.png')
plt.close()
I have done 24 hour measurement and results obtain contains around 1400 entries. Now I want to plot those results in such a way
That x axis represent my time and y axis the corresponding value.
My x axis should be divided into 24 sections each representing 1
hour.
My exact start time is 14:00 and end time is next day 14:00.
For more clarification I am adding a simple version of my data here below
And resulting Plot I am getting is this.
I look forward to your answers. Thank you.
If the time values go across midnight, you need to add a date part to the time value, so they can be plotted correctly as before and after midnight. At the very least, the time values for the first day should have a 0 before the decimal, e.g. 0.875 for 9 pm, and the values after midnight should have a 1 before the decimal, e.g. 1.125 for 1 am, so it falls on the next day and not the same day as the 9pm value.
Then plot an XY Scatter chart.
Work out what Excel's internal number (date/time value showing in General format) is for the desired X axis minimum, maximum and major/minor increments and format the x axis accordingly. Set the number format to hh:mm
Edit: For example: you want the minimum X axis value to be 24-Dec-2015 11 pm. Write that into a cell as a date/time. Format the cell to General. Then use the number you see in the format dialog for the X axis minimum.
If you want the major unit to be 1 hour, write the time value 1:00 into a cell and format it with general. Use that number in the dialog for Major.
Format the X axis labels to show time values, not dates.
greeting for every one,
I have data in excel file and i want to draw a plot in Matlab in which the Y axis represent the time with starting time in 10:45 for 24 hours i.e, from 10:00 am to the next day in 10:00 am. The x-axis represents the excel file data( the values of frequencies during 24 hours)
how to put the different times in the y axis showing the values of time in the formula of time(00:00 am/pm) using matlab?
if i use this code: ylim(subplot2,[1 24]) and xlim(subplot2,[170 230]) it will be plotted but the y-axis shows only the hours from 1 to 24 hours and i need the y-axis from 10:45 am(starting time) to(10:45)am in interval 24 hours
You can create custom tick labels by specifying tick strings with the command:
time_cells = {'10:45','11:45',...,'9:45','10:45'};
set(gca, 'YTickLabel', time_cells)
Where gca is the handle of your current plot (axes), and the time_cells is a cell array containing all your required tick labels (without the ellipse). It is probably easiest to generate this using a for-loop to create the numbers you want, and then num2str to convert to the strings you need.
I have data that looks like this:
1000 13 75.2
1000 21 79.21
1000 29 80.02
5000 29 87.9
5000 37 88.54
5000 45 88.56
10000 29 90.11
10000 37 90.79
10000 45 90.87
I want to use the first column as x axis labels, the second column as y axis labels and the third column as the z values. I want to display a surface in that manner. What is the best way to do this? I tried Excel but didn't really get anywhere. Does anyone have any suggestions for a tool to do this? Does anyone know how to do this in Excel?
Thanks
I ended up using matplotlib :)
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import cm
import matplotlib.pyplot as plt
import numpy as np
x = [1000,1000,1000,1000,1000,5000,5000,5000,5000,5000,10000,10000,10000,10000,10000]
y = [13,21,29,37,45,13,21,29,37,45,13,21,29,37,45]
z = [75.2,79.21,80.02,81.2,81.62,84.79,87.38,87.9,88.54,88.56,88.34,89.66,90.11,90.79,90.87]
fig = plt.figure()
ax = fig.gca(projection='3d')
ax.plot_trisurf(x, y, z, cmap=cm.jet, linewidth=0.2)
plt.show()
You really can't display 3 columns of data as a 'surface'. Only having one column of 'Z' data will give you a line in 3 dimensional space, not a surface (Or in the case of your data, 3 separate lines). For Excel to be able to work with this data, it needs to be formatted as shown below:
13 21 29 37 45
1000 75.2
1000 79.21
1000 80.02
5000 87.9
5000 88.54
5000 88.56
10000 90.11
10000 90.79
10000 90.87
Then, to get an actual surface, you would need to fill in all the missing cells with the appropriate Z-values. If you don't have those, then you are better off showing this as 3 separate 2D lines, because there isn't enough data for a surface.
The best 3D representation that Excel will give you of the above data is pretty confusing:
Representing this limited dataset as 2D data might be a better choice:
As a note for future reference, these types of questions usually do a little better on superuser.com.
You can use r libraries for 3 D plotting.
Steps are:
First create a data frame using data.frame() command.
Create a 3D plot by using scatterplot3D library.
Or You can also rotate your chart using rgl library by plot3d() command.
Alternately you can use plot3d() command from rcmdr library.
In MATLAB, you can use surf(), mesh() or surfl() command as per your requirement.
[http://in.mathworks.com/help/matlab/examples/creating-3-d-plots.html]
You also can use Gnuplot which is also available from gretl. Put your x y z data on a text file an insert the following
splot 'test.txt' using 1:2:3 with points palette pointsize 3 pointtype 7
Then you can set labels, etc. using
set xlabel "xxx" rotate parallel
set ylabel "yyy" rotate parallel
set zlabel "zzz" rotate parallel
set grid
show grid
unset key
Why not merge the rows that contain the same values?
-
13 21 29 37 45
1000] -75.2 -- 79.21 -- 80.02
5000] ---------------------87.9---88.54----88.56
10000] -------------------90.11--90.97----90.87
Excel can use that pretty well..