4. Vesuvius Viewsheds Visualised¶

Recent geothermal events have drawn our attention to historical depictions of the volcano Vesuvius. Let’s take a look at artworks in the Birmingham and V&A collections depicting Vesuvius, asking that all-important question, what historical viewsheds of Vesuvius are there ?

Note

Whilst pleasingly alliterative, the proper definition of a viewshed does not totally match up with the usage here, to show the estimated view of a painting when placed onto a map. Possibly a better definition would be line of sight, but that doesn’t have enough V’s in it.

Birmingham Museums Collections¶

A search for Naples shows 19 paintings, of which a handful give enough clues in their titles to give a very approximate location to place on a map.

V&A Collections¶

Searching for Vesuvius in the V&A collections reveals four watercolours we can estimate a location of.

Note

Please note, these are estimates of places based on very limited geographical knowledge by the writer of this notebook and is intended as a worked example for data exploration and should not be taken as a serious investigation into vulcanology or art history.

To make the notebook reproducable, we’ve created a small CSV file containing the information needed from both collections, giving 8 artworks to map.

This CSV contains the fields:

Holding Collection
Artwork Title
Artwork artist/maker
Artwork date of production (to the year)
Artwork description
Artwork link
Artwork thumbnail
Artwork viewshed lat/lon (estimated by the author of this notebook)
Artwork viewshed direction
Artwork viewshed confidence (marked by self)

import pandas as pd

vesuvius_df = pd.read_csv('vesuvius.csv')
vesuvius_df["Description"] = vesuvius_df["Description"].astype(str)
vesuvius_df

	Collection	Title	Maker	Description	Link	Thumbnail	Latitude	Longitude	Direction	Confidence
0	Birmingham Museums	The Bay of Naples from Capodimonte	William Paris	View from a vineyard between the Capodimonte a...	https://dams.birminghammuseums.org.uk/asset-ba...	NaN	40.858144	14.221832	180	70
1	Birmingham Museums	Naples From Sir William Hamilton's Villa	John Warwick Smith	nan	https://dams.birminghammuseums.org.uk/asset-ba...	NaN	40.815083	14.335572	225	50
2	Birmingham Museums	Temple of Venus, Bay of Baia	William Pars	nan	https://dams.birminghammuseums.org.uk/asset-ba...	NaN	40.818099	14.071142	225	70
3	Birmingham Museums	Near Sorrento	John Brett	nan	https://dams.birminghammuseums.org.uk/asset-ba...	NaN	40.634868	14.363853	225	75
4	V&A	Vesuvius	John Robert Cozens	Mount Vesuvius; Viewed from the land side; rui...	https://collections.vam.ac.uk/item/O1139859/ve...	NaN	40.800000	14.500000	270	25
5	V&A	Vesuvius and the Bay of Naples from Posilipo	John William Inchbold	Watercolour drawing	https://collections.vam.ac.uk/item/O187904/ves...	NaN	40.811328	14.183128	225	70
6	V&A	Street of the Tombs, Pompeii	Samuel Palmer"	'Street of the Tombs, Pompeii', showing the ru...	https://collections.vam.ac.uk/item/O126826/st...	NaN	40.746618	14.476173	270	50
7	V&A	Vesuvius	John Robert Cozens	Vesuvius; a slight volume of smoke issuing fro...	"https://collections.vam.ac.uk/item/O1023498/...	NaN	40.815000	14.335000	270	75

Viewshed Mapping¶

Now we can display these points on on a map. and the viewing angle (as a triangle) they very approximately have, either facing towards (or away) from Vesuvius. For this we are using leaflet and some hardcoded rules.

Note

For reasons unknown, the map below is sometimes showing more than once.

from ipyleaflet import Map, basemaps, basemap_to_tiles, GeoJSON, Polygon, Marker, Popup
from ipywidgets import HTML
from IPython.core import display
import numpy

m = Map(
    basemap=basemap_to_tiles(basemaps.OpenStreetMap.Mapnik),
    center=(40.8, 14.41),
    zoom=10
)

for index,obj in vesuvius_df.iterrows():

  origin_viewpoint = numpy.array([obj['Latitude'], obj['Longitude']])

  if obj['Collection'] == "Birmingham Museums":
        polygon_colour = "orange"
  else:
        polygon_colour = "red"

  # This really should be done by some angle calculations that are beyond me, hence just using 90,180,270...
  if obj['Direction'] == 90:
        viewshed_shape = [origin_viewpoint.tolist(), (origin_viewpoint + (-0.02,0.05)).tolist(),
                          (origin_viewpoint + (0.02, 0.05)).tolist()]
  elif obj['Direction'] == 180:
        viewshed_shape = [origin_viewpoint.tolist(), (origin_viewpoint + (-0.05,-0.02)).tolist(),
                          (origin_viewpoint + (-0.05, 0.02)).tolist()]
  elif obj['Direction'] == 270:
        viewshed_shape = [origin_viewpoint.tolist(), (origin_viewpoint + (-0.02,-0.05)).tolist(),
                          (origin_viewpoint + (0.02, -0.05)).tolist()]
  else:
        viewshed_shape = [origin_viewpoint.tolist(), (origin_viewpoint + (-0.02,0.05)).tolist(),
                          (origin_viewpoint + (0.02, 0.05)).tolist()]
        
  polygon = Polygon(
     locations= [ viewshed_shape  ],  color=polygon_colour,  fill_color=polygon_colour,
     stroke=True,  dash_array="4,1,1", weight=2)

  m.add_layer(polygon)

  marker = Marker(location=origin_viewpoint.tolist())
  m.add_layer(marker)

  message1 = HTML()
  message1.value = f'<a href="{obj["Link"]}">{obj["Title"]}</a> '
  if len(obj['Description']) > 0:
    message1.value += obj["Description"]

   # Popup with a given location on the map:
  popup = Popup(
    location=origin_viewpoint.tolist(), child=message1, close_button=True,
    auto_close=True, close_on_escape_key=False
  )
      
#  m.add_layer(popup)
  marker.popup = message1

m.layout.width = '100%'
m.layout.height = '500px'

m.save('vesuvius-viewsheds.html', title='Vesuvius Viewsheds')
display.HTML('vesuvius-viewsheds.html')

Vesuvius Viewsheds

Future Enhancements¶

Enable timeslide plugin in leaflet (needs to be ported to Jupyter ipyleaflet) to enable showing viewsheds over time
Supress nan values appearing in popups
Have more data assembled by someone who has real knowledge of this topic.
Do this properly in 3D as the Georgia O’Keeffe Museum have done

Data Explorations

4. Vesuvius Viewsheds Visualised¶

Birmingham Museums Collections¶

V&A Collections¶

Viewshed Mapping¶

Future Enhancements¶