Project "Coastal Opportunity Analysis" analyzes global shark attack data to identify optimal geographic expansion opportunities for our water sports equipment business.

The core objective is to determine where high ocean activity and low fatality rates create strong revenue potential and high Customer Lifetime Value (CLV).

If we expand into markets with:

• High shark attack frequency (proxy for high ocean activity)
• Low fatality rates (indicating manageable risk and sustained participation)

Then sales will increase due to:

• Strong water sports participation
• Ongoing customer engagement
• Higher repeat purchase potential (CLV)

Note: The United States emerges as a primary market candidate under this framework.

This project uses the Shark Attack dataset, which includes global records of shark–human interactions.

Key variables that we considered for analysis are the following:

• Country/location
• Activity type (e.g., surfing, swimming, diving)
• Injury severity
• Fatal vs. non-fatal outcome
• Species involved (where available)

Note: The dataset enables geographic, temporal, and risk-level analysis.

To ensure accuracy and consistency, the following preprocessing steps were applied:

• Standardized country/location names

  df.columns = df.columns.str.strip().str.lower().str.replace(" ", "_")
  

• Normalized injury outcomes into binary fatal/non-fatal classification

  fatal_replacements = {"Y x 2": "Y", " N": "N", "Nq":"N", "M": np.nan, "F": np.nan, "2017": np.nan, "UNKNOWN": np.nan}
  

df["fatal_y/n"] = (df["fatal_y/n"].replace(fatal_replacements).str.strip().str.upper()) print(df["fatal_y/n"].value_counts(dropna = False))

• Grouped unstructured data into standardized categories

  fatality_rate = df.groupby('country')['fatal_numeric'].mean() * 100
  
  or
  
  def clean_activity(row):
      activity = row['Activity']  # Récupérer l'activité de la ligne
      if pd.isna(activity):  # Si la valeur est NaN, la laisser inchangée
          return row['activity_clean']
      if "swimming" in activity.lower():
          return "Swimming"
      elif "fishing" in activity.lower():
          return "Fishing"
      elif "surfing" in activity.lower():
          return "Surfing"
      elif "wading" in activity.lower():
          return "Wading"
      elif "Diving" in activity.lower():
          return "Diving"
      else:
          return row['activity_clean']
  

• Removed records with missing data

  fatality_rate = fatality_rate.dropna()
  
  or
  df = df.replace("?","")
  

• Cleaned text fields (lowercasing, trimming, removing special characters)

  df.loc[:, "country"] = df["country"].str.replace(r"[^a-z\s]", "", regex=True)
  

Note: These steps improved reliability for cross-country comparison and fatality rate analysis.

1. Calculated shark incident frequency by country
2. Computed fatality rates per country
3. Identified high-incident, low-fatality markets
4. Evaluated expansion attractiveness based on:

• Activity volume (proxy for demand)
• Risk sustainability
• Potential CLV impact

Markets with high shark activity but low fatality rates indicate:

• Strong ocean sports culture
• Sustained consumer participation
• Repeat equipment purchase potential
• Long-term revenue growth opportunity

Note: The United States ranks highly under this strategic model.

1. /data

• https://www.sharkattackfile.net/spreadsheets/GSAF5.xls

2. /notebooks -Shark Project.ipynb

3. README.md

• Python
• Pandas
• NumPy
• Jupyter Notebook
• Google Cloud
• Excel

Anne Leschallier de Lisle, Beatriz Fernandes, Francisca Andrade, Ofelia Akopian

Presentation

Voir le Google Slides