CODEWITHSUNDEEP
pythonpandasnumpymatplotlib
jerickson759
jerickson759

Reputation: 21

Matplotlib Filled In Step Graph for Weights

I have a DataFrame that contains the history of weight values for N items. The sum of the weights for all items is 1. The date is when the weights were set, and they stay the same until they are set again.

Here is the Dataframe.

            item2     item3     item4     item1
2018-03-26  0.333331  0.333331  0.000008  0.333331
2018-03-29  0.097470  0.249645  0.250353  0.402532
2018-04-01  0.329531  0.080273  0.356974  0.233221
2018-04-04  0.279514  0.048366  0.288980  0.383140
2018-04-07  0.289597  0.082032  0.201930  0.426442
2018-04-10  0.258850  0.107432  0.280754  0.352964
2018-04-13  0.030390  0.172697  0.486147  0.310766
2018-04-16  0.019628  0.095833  0.433445  0.451094
2018-04-19  0.054276  0.270346  0.351813  0.323565
2018-04-22  0.075713  0.089441  0.385883  0.448963
2018-04-25  0.109152  0.071784  0.536841  0.282223
2018-04-28  0.172047  0.108976  0.360786  0.358192
2018-05-01  0.059381  0.072429  0.434918  0.433272

I want to visualize how the weights change over time. I came up with this.

#weight_history is the DataFrame
tot = np.zeros(len(weight_history))
for weight in weight_history:
    plt.step(weight_history[weight].index, weight_history[weight].values+tot,where='post',label=weight)
    tot += weight_history[weight].values
plt.set_ylabel("Weights")
plt.legend()

Which produces this:

enter image description here

This is very close to what I want, but it is hard to visualize how much weight each item has.

How can I fill in the area under each line?

So filled blue from 0 to the blue line, filled orange from the blue line to the orange line, filled green from the orange line to the green line, and filled red from the green line to the red line.

Is there an alternative way to do this?

Upvotes: 1

Views: 504

Answers (2)

jerickson759
jerickson759

Reputation: 21

I figured it out, fill_between has a step option.

tot = np.zeros(len(weight_history))
for weight in weight_history:
    axes[1].fill_between(x=weight_history.index, y1=weight_history[weight].values+tot,y2=tot,step='post',label=weight)
    tot += weight_history[weight].values
axes[1].set_ylabel("Weights")
axes[1].legend()

Upvotes: 1

Richard Rublev
Richard Rublev

Reputation: 8162

Try something like this

plt.fill_between(x, item2, item3, color='blue')
plt.show()

Upvotes: 0

Related Questions