# Which graph will be appropriate for the visualization task?

I have some terminal charging values for US and CHINA comes in a pandas DataFrame like the following,

       value country
0      550.0     USA
1      820.0   CHINA
2      835.0   CHINA
3      600.0   USA
4      775.0   CHINA
5      785.0   USA
6      790.0   USA


This is the sample data and I have in total 5K+ entries. The data is cleared for the outliers and needs to be visualized. What kind of visualization can I use to plot my data meaningfully?

• What specific visualization task do you need? Outliers detection or something else? – Icyblade Mar 8 '17 at 19:00
• My instruction is The list of values per country should be visualised in a meaningful way, identifying the amount of good/bad charges present in the system. I have cleared the outliers earlier and this data is good. I thought the comparable visualization per country would be great while I can manually calculate the percentage of the outliers per country and store using some method. – Chaklader Asfak Arefe Mar 8 '17 at 19:04
• asymmetric beanplot maybe? example + statsmodels implementation (use 'side' argument with 'left' and 'right') – oW_ Mar 8 '17 at 20:07

Maybe you can try something like this?

df.hist(by='country', bins=50)


If you want to plot these histograms in the same figure, checkout this.

## References

http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.hist.html

If you need to be able to identify good and bad charge times then the points themselves are important.

I would have my x-axis be my country names and then simply plot the charge times on the y-axis as scattered points.

This will give you scattered data in the y-direction digitized for each country.

I think you will be able to extract a lot of meaningful information from this plot for comparing different countries such as: max, min, approx. mean, etc..

It depends on what is your objective.

Since you have so many values, you may want to plot some summary data, so you can use either two Distplots in a FacetGrid to have the two countries side by side or even better a Boxplot.

If the order of the appearance of the values matters you can try plot them as two different lines with the values on the y-axis by using FactorPlot.

(All the mentioned plots above are part of the Seaborn library)