index.qmd

---
title: "IGSN Resolution Metrics"
subtitle: "Summary for 2022-07-01 ... 2022-08-31."
jupyter: python3
execute:
  echo: false
format:
  html:
    toc: false
    theme:
      - cosmo
---

Some basic statistics for use of the igsn.org resolver service use during the months of July and August, 2022.

```{python}
import os
import sqlite3
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import IPython.display
import itables

itables.init_notebook_mode(all_interactive=True)

ANALYSIS_DIR = "analysis"
ANALYSIS_DB = os.path.join(os.path.abspath(ANALYSIS_DIR),"logs.sqlite3")
CN = sqlite3.connect(ANALYSIS_DB)

# Load the country names and codes
ccodes = pd.read_csv("geo/country_codes.csv", quotechar='"')
ccodes.rename(mapper={"Alpha-2 code":"country_code", "Alpha-3 code":"iso_a3"}, axis=1, inplace=True)
ccodes = ccodes.apply(lambda x: x.str.strip(' "'))
```

## Total redirects per day

The total number of redirects to `hdl.handle.net` handled by igsn.org per day. Such requests arise from a user filling in the form on the igsn.org site or directly by constructing the URL.

```{python}
sql = '''select count(*) as n, 
  cast(strftime('%d', t) as integer) as Day, 
  cast(strftime('%m', t) as integer) as m from logs 
  where t>='2022-07-01T00:00:00' 
  and redirect_url like 'http%hdl.handle.net%' 
  group by m,Day 
  order by m,Day asc;
'''
df = pd.read_sql_query(sql, CN)

fig, ax = plt.subplots(figsize=(8,3))
for labels, dfi in df.groupby("m"):
  dfi.plot(ax=ax, x='Day', y='n', label=labels);
a = ax.legend(title="Month");
#IPython.display.display(fig)
```

## Total redirects per month by country

The country of origin for redirect requests issued by igsn.org. Column `n` is the total requests for the month for the specified country.

::: {.panel-tabset}

## July
```{python}
month_map = {
  7:"July",
  8: "August"
}

sql = '''select count(*) as n, 
country_code, 
cast(strftime('%m', t) as integer) as m from logs 
where t>='2022-07-01T00:00:00' 
and redirect_url like 'http%hdl.handle.net%'
group by m, country_code 
order by m asc, n desc'''
df = pd.read_sql_query(sql, CN)
df_merged = pd.merge(
  df, 
  ccodes, 
  how="left", 
  on=["country_code"]
)
#df_merged.style.hide(axis=0)
m = 7
dfm = df_merged[df_merged['m']==m].loc[:, ['n','Country', 'country_code']].drop_duplicates(subset=['country_code'])
_ = dfm.style.hide(axis=0)
itables.show(
  dfm, 
  scrollY="480px", 
  scrollCollapse=True, 
  paging=False,
  dom='tpr'
)
```

## August

```{python}
m = 8
dfm = df_merged[df_merged['m']==m].loc[:, ['n','Country', 'country_code']].drop_duplicates(subset=['country_code'])
_ = dfm.style.hide(axis=0)
itables.show(
  dfm, 
  scrollY="480px", 
  scrollCollapse=True, 
  paging=False,
  dom='tpr'
)
```

:::

Same content as above, shown as a map.

```{python}
#| column: page
import geopandas
import matplotlib.colors
import numpy as np

def truncate_colormap(cmap, minval=0.0, maxval=1.0, n=100):
    new_cmap = matplotlib.colors.LinearSegmentedColormap.from_list(
        'trunc({n},{a:.2f},{b:.2f})'.format(n=cmap.name, a=minval, b=maxval),
        cmap(np.linspace(minval, maxval, n)))
    return new_cmap

world = geopandas.read_file(
    geopandas.datasets.get_path('naturalearth_lowres')
)
ccodes = pd.read_csv("geo/country_codes.csv", quotechar='"')
ccodes.rename(mapper={"Alpha-2 code":"country_code", "Alpha-3 code":"iso_a3"}, axis=1, inplace=True)
ccodes = ccodes.apply(lambda x: x.str.strip(' "'))
merged = pd.merge(df.set_index("country_code"), ccodes.set_index("country_code"), how="left", on=["country_code"])
for _month in [7,8]:
  wmerged = pd.merge(
    world.set_index("iso_a3"), 
    merged[(merged.m==_month)].set_index("iso_a3"), 
    how="left", 
    on=["iso_a3"]
  )
  fig, ax = plt.subplots(figsize=(12, 5));
  _ = ax.axis('off');
  _ = ax.set_title(
    f'igsn.org resolve request origins for {month_map[_month]} 2022', 
    fontdict={'fontsize': '10', 'fontweight' : '3'}
  );
  wmerged.plot(column='n', 
    cmap='viridis', 
    linewidth=0.9, 
    ax=ax, 
    edgecolor='1',
    legend=True,
    norm=matplotlib.colors.LogNorm(vmin=wmerged.n.min(), vmax=wmerged.n.max()),
    missing_kwds={
      "color": "lightgrey",
      "label": "Missing values",
    })
```

## User agents used for resolve requests

Total requests for July and August 2022 grouped by browser family and brand of the device issuing the resolve requests.

```{python}
sql = '''select count(*) as n, 
browser_family, 
device_brand
from logs 
where t>='2022-07-01T00:00:00' 
and redirect_url like 'http%hdl.handle.net%'
group by browser_family, device_brand
order by n desc'''
df = pd.read_sql_query(sql, CN)
itables.show(df,
  scrollY="480px", 
  scrollCollapse=True, 
  paging=False,
  dom='tpr'
)
```