Skip to content
Snippets Groups Projects
Commit db521fa7 authored by Chandrasekhar Ramakrishnan's avatar Chandrasekhar Ramakrishnan
Browse files

fix: Use geodata for centroids in Dashboard

parent 263e5a18
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
notebooks/Dashboard.ipynb
+ 46
− 30
View file @ db521fa7
%% Cell type:code id: tags:
``` python
import pandas as pd
import os
from IPython.display import display, HTML, Markdown
```
%% Cell type:code id: tags:parameters
``` python
ts_folder = "../data/covid-19_jhu-csse/"
rates_folder = "../data/covid-19_rates/"
geodata_path = "../data/geodata/geo_data.csv"
out_folder = None
PAPERMILL_OUTPUT_PATH = None
```
%% Cell type:code id: tags:
``` python
# Read in the data
```
%% Cell type:code id: tags:
``` python
def read_jhu_covid_df(name):
filename = os.path.join(ts_folder, f"time_series_19-covid-{name}.csv")
df = pd.read_csv(filename)
df = df.set_index(['Province/State', 'Country/Region', 'Lat', 'Long'])
df.columns = pd.to_datetime(df.columns)
return df
jhu_frames_map = {
"confirmed": read_jhu_covid_df("Confirmed"),
"deaths": read_jhu_covid_df("Deaths"),
"recovered": read_jhu_covid_df("Recovered")
}
```
%% Cell type:code id: tags:
``` python
def read_rates_covid_df(name):
filename = os.path.join(rates_folder, f"ts_rates_19-covid-{name}.csv")
df = pd.read_csv(filename).drop("Unnamed: 0", axis=1)
df = df.set_index(['Country/Region'])
df.columns = pd.to_datetime(df.columns)
return df
rates_frames_map = {
"confirmed": read_rates_covid_df("confirmed"),
"deaths": read_rates_covid_df("deaths"),
"recovered": read_rates_covid_df("recovered")
}
```
%% Cell type:code id: tags:
``` python
# Compile data needed for the visualizations
geodata_df = pd.read_csv(geodata_path)
geodata_df = geodata_df.drop('Unnamed: 0', axis=1)
geodata_df = geodata_df.rename({'name_jhu':'Country/Region'}, axis=1)
geodata_df = geodata_df.set_index('Country/Region')
```
%% Cell type:code id: tags:
``` python
# Compute geospatial coordinates
country_coords_df = jhu_frames_map['confirmed'].reset_index([2,3])[['Lat', 'Long']]
country_coords_df = country_coords_df.groupby(level='Country/Region').mean()
```
# Identify countries with 100 or more cases
def latest_jhu_country_ser(name):
return jhu_frames_map[name].iloc[:,-1].groupby(level='Country/Region').sum()
%% Cell type:code id: tags:
``` python
# Identify countries with 100 or more cases
case_count_ser = jhu_frames_map['confirmed'].iloc[:,-1].groupby(level='Country/Region').sum()
case_count_ser = latest_jhu_country_ser('confirmed')
countries_over_thresh = case_count_ser[case_count_ser > 99].index
```
%% Cell type:markdown id: tags:
# Questions About COVID-19 and Its Spread
These plots should be taken with a large grain of salt. I am not an epidemiologist, so the analyses shown here are completely naive. There are large discrepencies in the data from different countries for a variety of reasons (rates of testing, demographics, etc.) so that make direct comparisons inaccurate. Nonetheless, I think there is a lot of interesting information in this data.
Understanding the spread, distribution, and deadliness of COVID-19 is difficult, despite the data available about it. Differences in rates of testing, quality of data, demographics, etc. make it difficult to compare data between countries.
All this needs to be considered when looking at the plots below. But despite those caveats, I found it helpful to plot the raw data, even though direct comparisons between countries might not be inaccurate.
%% Cell type:code id: tags:
``` python
data_ts = jhu_frames_map['confirmed'].iloc[:,-1].name.strftime("%b %d %Y")
display(HTML(f"<em>Data up to {data_ts}</em>"))
display(HTML(f"<em>Data up to {data_ts}; countries with 100 or more confirmed cases.</em>"))
```
%% Cell type:markdown id: tags:
## How are cases per 100,000 distributed geographically?
%% Cell type:code id: tags:
``` python
import altair as alt
from vega_datasets import data
```
%% Cell type:code id: tags:
``` python
def latest_rates_ser(name):
return rates_frames_map[name].iloc[:,-1]
# Compile the basic df
map_df = pd.concat([
rates_frames_map['confirmed'].iloc[:,-1],
rates_frames_map['deaths'].iloc[:,-1],
rates_frames_map['recovered'].iloc[:,-1],
country_coords_df], axis=1)
latest_rates_ser('confirmed'),
latest_rates_ser('deaths'),
latest_rates_ser('recovered')], axis=1)
nominal_df = pd.concat([
latest_jhu_country_ser('confirmed'),
latest_jhu_country_ser('deaths'),
latest_jhu_country_ser('recovered')], axis=1)
map_df = pd.concat([map_df, nominal_df, geodata_df[['Longitude', 'Latitude']]], axis=1)
# Restrict to countries with 100 or more cases
map_df = map_df.loc[countries_over_thresh].dropna()
map_df = map_df.reset_index()
map_df.columns = ['Country/Region', 'Confirmed', 'Deaths', 'Recovered', 'Lat', 'Long']
map_df.columns = ['Country/Region',
'Confirmed/100k', 'Deaths/100k', 'Recovered/100k',
'Confirmed', 'Deaths', 'Recovered',
'Long', 'Lat']
```
%% Cell type:code id: tags:
``` python
def map_of_variable(map_df, variable):
def map_of_variable(map_df, variable, title):
# Data generators for the background
sphere = alt.sphere()
graticule = alt.graticule()
# Source of land data
source = alt.topo_feature(data.world_110m.url, 'countries')
# Layering and configuring the components
p = alt.layer(
alt.Chart(sphere).mark_geoshape(fill='#cae6ef'),
alt.Chart(graticule).mark_geoshape(stroke='white', strokeWidth=0.5),
alt.Chart(source).mark_geoshape(fill='#dddddd', stroke='#aaaaaa'),
alt.Chart(map_df).mark_circle(opacity=0.6).encode(
longitude='Long:Q',
latitude='Lat:Q',
size=alt.Size(f'{variable}:Q', title="Cases"),
color=alt.value('steelblue'),
tooltip=["Country/Region:N", "Confirmed:Q", "Deaths:Q", "Recovered:Q"]
tooltip=["Country/Region:N",
"Confirmed:Q", "Deaths:Q", "Recovered:Q",
"Confirmed/100k:Q", "Deaths/100k:Q", "Recovered/100k:Q"]
)
).project(
'naturalEarth1'
).properties(width=600, height=400, title=f"{variable} cases per 100,000"
).properties(width=600, height=400, title=f"{title} cases per 100k inhabitants"
).configure_view(stroke=None)
return p
```
%% Cell type:code id: tags:
``` python
display(map_of_variable(map_df, 'Confirmed'))
display(map_of_variable(map_df, 'Confirmed/100k', 'Confirmed'))
display(HTML('''
<p style="font-size: smaller">Data Source:
<a href="https://github.com/CSSEGISandData/COVID-19">JHU CSSE</a> and
<a href="https://data.worldbank.org/indicator/SP.POP.TOTL">World Bank</a>
<p style="font-size: smaller">Data Sources:
<a href="https://github.com/CSSEGISandData/COVID-19">JHU CSSE</a>,
<a href="https://data.worldbank.org/indicator/SP.POP.TOTL">World Bank</a>,
<a href="https://worldmap.harvard.edu/data/geonode:country_centroids_az8">Harvard Worldmap</a>
</p>'''))
```
%% Cell type:code id: tags:
``` python
bars = alt.Chart(map_df).mark_bar().encode(
x='Confirmed:Q',
y=alt.Y("Country/Region:N", sort='-x')
x='Confirmed/100k:Q',
y=alt.Y("Country/Region:N", sort='-x'),
tooltip=["Country/Region:N",
"Confirmed:Q", "Deaths:Q", "Recovered:Q",
"Confirmed/100k:Q", "Deaths/100k:Q", "Recovered/100k:Q"]
)
text = bars.mark_text(
align='left',
baseline='middle',
dx=3 # Nudges text to right so it doesn't appear on top of the bar
).encode(
text=alt.Text('Confirmed:Q', format=".3")
text=alt.Text('Confirmed/100k:Q', format=".3")
)
(bars + text).properties(height=900)
```
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment

This GitLab is a part of renkulab.io