---
jupytext:
  text_representation:
    extension: .md
    format_name: myst
  encoding: '# -*- coding: utf-8 -*-'
kernelspec:
  display_name: Python 3 (ipykernel)
  language: python
  name: python3
language_info:
  name: python
  nbconvert_exporter: python
  pygments_lexer: ipython3
nbhosting:
  title: basic numpy
---

# data loading

```{code-cell} ipython3
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
```

Let's load a dataset on rain precipitations on Seattle on 2014

```{code-cell} ipython3
:lines_to_next_cell: 2

# we download the file from Internet and save it
# easiest way, we can pass a URL to read_csv (or a local file)
URL = "http://www-sop.inria.fr/members/Arnaud.Legout/formationPython/Exos/Seattle2014.csv"

# don't worry, we will come back to this line when we will talk about pandas.
# for now it just load a ndarray
rainfall = pd.read_csv(URL)["PRCP"].to_numpy()

# other solution to get the remote file with urllib
# from urllib.request import urlopen
# with open("Seattle2014.csv", "w", encoding='utf-8') as f:
#    with urlopen(URL) as u:
#        f.write(u.read().decode('utf-8'))

# we extract with pandas the precipitation column
# rainfall is an array of precipitation per day 
# for each day of 2014
# rainfall = pd.read_csv('Seattle2014.csv')['PRCP'].to_numpy()
```

## Let's visualize

+++

**[assignement]**: plot the amount of rain (in mm) over time; make sure you put a proper label on both axes, and on the global figure

```{code-cell} ipython3
# your code here
```

## Let's answer the following questions

+++

**What is the shape and dype of the ndarray?**

```{code-cell} ipython3
# your code here
```

**How many rainy days?**

```{code-cell} ipython3
# your code here
```

**Average precipitation on the year?**

```{code-cell} ipython3
# your code here
```

**Average precipitation on the rainy days?**

```{code-cell} ipython3
# your code here
```

**Mean precipitation on January?**

```{code-cell} ipython3
# your code here
```

**Mean precipitation on January on the rainy days?**

```{code-cell} ipython3
# your code here
```

# A transition to pandas

```{code-cell} ipython3
# But in practice we don’t do that. Here is what we do…
# We start to convert to a pandas Series
s = pd.Series(rainfall)

# then we convert the index to the real dates
s.index = pd.to_datetime(s.index, unit='D',
                         origin=pd.Timestamp('1/1/2004'))

# possibly resample per month to get the total monthly rain
s = s.resample('m').max()
```

```{code-cell} ipython3
# then plot

%matplotlib ipympl

s.plot.bar()
plt.xlabel('month')
plt.ylabel('mm')
plt.title('Rainy days in 2014 at Seattle')
fig = plt.gcf()
fig.autofmt_xdate()
# plt.show() # if in a terminal
```

***