Intepretation of HLL data:
Comparison, Interactive Exploration, Benchmark Data

Table of Contents

• 1  Introduction
• 2  Preparations
  • 2.1  Parameters
  • 2.2  Load additional dependencies
• 3  Interactive Map with Holoviews/ Geoviews
  • 3.1  Load & plot pickled dataframe
  • 3.2  Prepare interactive mapping
• 4  Compare RAW and HLL
• 5  Working with the Benchmark data: Intersection Example
  • 5.1  Load Benchmark data
  • 5.2  Union hll sets for Countries UK, DE and FR
    • 5.2.1  Selection of grid cells based on country geometry
    • 5.2.2  Intersection with grid
    • 5.2.3  Plot preview of selected grid cells (bins)
  • 5.3  Union of hll sets
  • 5.4  Calculate intersection (common visitors)
  • 5.5  Illustrate intersection (Venn diagram)
  • 5.6  Get accurate intersection counts (raw)
    • 5.6.1  Prepare intersection with countries geometry
    • 5.6.2  Collect raw metrics
    • 5.6.3  Update Venn Diagram
  • 5.7  Final plot (Map & Venn Diagram)
• 6  Close DB connection & Create notebook HTML
• 7  Create release file

Introduction

This is the last notebook in a series of four notebooks:

• 1) the Preparations (01_preparations.ipynb) Basic preparations for processing YFCC100m, explains basic concepts and tools for working with the lbsn data
• 2) the RAW Notebook (02_yfcc_gridagg_raw.ipynb) demonstrates how a typical grid-based visualization looks like when using the raw lbsn structure and
• 3) the HLL Notebook (03_yfcc_gridagg_hll.ipynb) demonstrates the same visualization using the privacy-aware hll lbsn structure
• 4) the Interpretation (04_interpretation.ipynb) illustrates how to create interactive graphics for comparison of raw and hll results

In this notebook, we'll illustrate how to make further analyses based on the published hll benchmark dataset.

• create interactive map with geoviews, adapt visuals, styling and legend
• combine results from raw and hll into interactive map (on hover)
• store interactive map as standalone HTML
• exporting benchmark data
• intersecting hll sets for frequentation analysis

Preparations

Parameters

Load global settings and methods from 03_yfcc_gridagg_hll notebook. This will also load all parameters and methods defined in 02_yfcc_gridagg_raw.

import sys
from pathlib import Path
module_path = str(Path.cwd().parents[0] / "py")
if module_path not in sys.path:
    sys.path.append(module_path)

from _03_yfcc_gridagg_hll import *

Load additional dependencies

Load all dependencies at once, as a means to verify that everything required to run this notebook is available.

import geoviews as gv
import holoviews as hv
import geoviews.feature as gf
import pickle
import ipywidgets as widgets
from ipywidgets.embed import embed_minimal_html
from rtree import index
from geopandas.tools import sjoin
from geoviews import opts
from bokeh.models import HoverTool, FixedTicker
from matplotlib_venn import venn3, venn3_circles

Enable bokeh

preparations.init_imports()