Label points in dataframe Python

Question

I would like to label data points in pandas with x axis value. And I am trying to apply this solution into my code: Annotate data points while plotting from Pandas DataFrame

I'm getting an error saying:

AttributeError: 'PathCollection' object has no attribute 'text'

Here's my code:

def draw_scatter_plot(xaxis, yaxis, title, xaxis_label, yaxis_label, save_filename, color, figsize=(9, 7), dpi=100):
    fig = plt.figure(figsize=figsize, dpi=dpi)

    ax = plt.scatter(xaxis, yaxis, c=color)

    plt.xlabel(xaxis_label)
    plt.ylabel(yaxis_label)

    label_point(xaxis, yaxis, xaxis, ax)

    plt.title(title)

    fig.savefig(save_filename, dpi=100)

# label code from https://stackoverflow.com/questions/15910019/annotate-data-points-while-plotting-from-pandas-dataframe/15911372#15911372

def label_point(x, y, val, ax):
    a = pd.concat({'x': x, 'y': y, 'val': val}, axis=1)
    for i, point in a.iterrows():
        ax.text(point['x'], point['y'], str(point['x']))

Any advice on this problem?

Pandas' plot returns an axes-object [docs](https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.plot.html). You use plt.scatter, which returns a paths-object [docs](https://matplotlib.org/api/pyplot_api.html?highlight=scatter#matplotlib.pyplot.scatter). — sascha, Aug 14 '17 at 18:20
Create an axes-object manually, use it to call scatter, then pass it to your label-func. I'm not sure what's best practice, but ```f, ax = plt.subplots(1)``` followed by ```ax.scatter()``` (maybe ```ax[0]```)) might work. (although it looks silly because of subplot**s**) — sascha, Aug 14 '17 at 18:25
Then you should probably show your modified code. **Edit:** or use MaxU's more clever approach! — sascha, Aug 14 '17 at 18:31

score 3 · Answer 1 · answered Aug 14 '17 at 18:32

Consider the following demo:

In [6]: df = pd.DataFrame(np.random.randint(100, size=(10, 2)), columns=list('xy'))

In [7]: df
Out[7]:
    x   y
0  44  13
1  69  53
2  52  80
3  72  64
4  66  42
5  96  33
6  31  13
7  61  81
8  98  63
9  21  95

In [8]: ax = df.plot.scatter(x='x', y='y')

In [9]: df.apply(lambda r: ax.annotate(r['x'].astype(str)+'|'+r['y'].astype(str), 
                                       (r.x*1.02, r.y*1.02)), axis=1)
Out[9]:
0    Annotation(44,13,'44|13')
1    Annotation(69,53,'69|53')
2    Annotation(52,80,'52|80')
3    Annotation(72,64,'72|64')
4    Annotation(66,42,'66|42')
5    Annotation(96,33,'96|33')
6    Annotation(31,13,'31|13')
7    Annotation(61,81,'61|81')
8    Annotation(98,63,'98|63')
9    Annotation(21,95,'21|95')
dtype: object

Result:

Very nice combination of matplotlib and pandas! – sascha Aug 14 '17 at 18:36 — sascha, Aug 14 '17 at 18:36

score 0 · Answer 2 · answered Aug 15 '17 at 09:35

The problem appears because you give the return of plt.scatter the name ax. This is confusing, because it is not an axes, but a 'PathCollection' (as the error tells you).

Replace the first two lines

fig = plt.figure(figsize=figsize, dpi=dpi)
ax = plt.scatter(xaxis, yaxis, c=color)

by

fig, ax = plt.subplots(figsize=figsize, dpi=dpi)
ax.scatter(xaxis, yaxis, c=color)

and keep the rest of the code the same.

Label points in dataframe Python

2 Answers2