In relation this post, I have acquired the raw disdrometer data which includes multiple pairs of Drop Size (0.062 mm, 0.187 mm etc) and Counts of raindrops (32 bins). I would like to create a time series heatmap of the disdrometer data in which the independent variables are Time and Drop Size, while the dependent variable is "log N(d)". How can I efficiently create this heatmap time series in Python? I did try to do this with the initial code, but I didn't manage to reproduce the sample figure in the previous post, correctly (the visualization). is there a way to adjust the x and y limits, the x and y tickmarks/labels as well? (as they were quite incorrect). To compute for N(d), the formula we use is "Sum of Counts in each time stamp" / 0.54 * 300 * 0.125 I did try my best to do this with the initial code, but I didn't manage to reproduce the sample figure in the previous post, correctly (the visualization). Like is there a way to adjust/correct the x and y limits? Also the tickmarks/labels in x and y axis were quite incorrect. The sample disdrometer data can be acquired through here. I'll appreciate a helping hand to this matter. import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns # File paths for all datasets all_file_paths = { "0.062 mm": "07292023/0.062 mm.csv", "0.187 mm": "07292023/0.187 mm.csv", "0.312 mm": "07292023/0.312 mm.csv", "0.437 mm": "07292023/0.437 mm.csv", "0.562 mm": "07292023/0.562 mm.csv", "0.687 mm": "07292023/0.687 mm.csv", "0.812 mm": "07292023/0.812 mm.csv", "0.937 mm": "07292023/0.937 mm.csv", "1.062 mm": "07292023/1.062 mm.csv", "1.187 mm": "07292023/1.187 mm.csv", "1.375 mm": "07292023/1.375 mm.csv", "1.625 mm": "07292023/1.625 mm.csv", "1.875 mm": "07292023/1.875 mm.csv", "2.125 mm": "07292023/2.125 mm.csv", "2.375 mm": "07292023/2.375 mm.csv", "3.25 mm": "07292023/3.25 mm.csv", "3.75 mm": "07292023/3.75 mm.csv", "4.25 mm": "07292023/4.25 mm.csv", "4.75 mm": "07292023/4.75 mm.csv", "5.5 mm": "07292023/5.5 mm.csv", "6.5 mm": "07292023/6.5 mm.csv", "7.5 mm": "07292023/7.5 mm.csv", "8.5 mm": "07292023/8.5 mm.csv", } # Constants for N(D) calculation factor = 0.54 * 300 * 0.125 # Read and process all datasets combined_data = [] for size, path in all_file_paths.items(): df = pd.read_csv(path) df["Size"] = float(size.split()[0]) # Ensure size is numeric time_col = df["Time stamp"] # Keep original time format nd_values = df.iloc[:, 2:].astype(float) / factor # Compute N(D) # Sum bins within 1 mm intervals size_bin = int(df["Size"].iloc[0] // 0.25) sum_nd = nd_values.sum(axis=1) # Sum all bins in this interval log_nd = np.log10(sum_nd.replace(0, np.nan)) # Compute log N(D) # Create DataFrame with relevant information result_df = pd.DataFrame({ "Time stamp": time_col, "Size_bin": size_bin, "log N(D)": log_nd }) combined_data.append(result_df) # Combine into a single DataFrame heatmap_data = pd.concat(combined_data).reset_index(drop=True) # Filter for log N(D) > 0 #heatmap_data = heatmap_data[heatmap_data["log N(D)"] > 0] # Aggregate log N(D) by size bin and time heatmap_binned = heatmap_data.groupby(["Time stamp", "Size_bin"])['log N(D)'].mean().reset_index() # Pivot for heatmap heatmap_pivot = heatmap_binned.pivot(index="Size_bin", columns="Time stamp", values="log N(D)") # Plot heatmap using axes fig, ax = plt.subplots(figsize=(12, 6)) sns.heatmap(heatmap_pivot, cmap="turbo", robust=True, vmin=0, vmax=2, ax=ax, cbar_kws={'pad': 0.1}) ax.set_xlabel("Time Stamp (LST)", fontsize=15) ax.set_ylabel("Drop Size (mm)", fontsize=15) #ax.set_title("Raindrop Size Distribution Heatmap (Summed Bins)") ax.invert_yaxis() # Invert y-axis to start from lowest to highest size # Set y-axis tick labels to represent bin intervals ax.set_yticks(np.arange(0, heatmap_pivot.shape[0], 1)) ax.set_yticklabels(np.arange(0, heatmap_pivot.shape[0], 1)) custom_labels = np.arange(1, heatmap_pivot.shape[0] + 1) ax.set_yticklabels(custom_labels) #ax.minorticks_on() ax.tick_params(axis='x', which='major', labelsize=15, length=10, width=2, top=True, right=True, direction='in') ax.tick_params(axis='y', which='major', labelsize=15, length=10, width=2, top=True, right=True, direction='in', rotation=0) #ax.tick_params(which='minor', length=7, width=1.5, top=True, right=True, direction='in') # Customize tick parameters for colorbar cbar = ax.collections[0].colorbar cbar.set_label('log N(D)', fontsize=15) cbar.ax.minorticks_on() cbar.ax.tick_params(which='major', labelsize=15, length=10, width=2, top=True, right=True, direction='in') cbar.ax.tick_params(which='minor', length=5, width=1.5, top=True, right=True, direction='in') ax.grid(ls='--', alpha=0.35) # Add border to the colorbar cbar.outline.set_edgecolor('black') cbar.outline.set_linewidth(2) # Add border to the plot plt.gca().spines['top'].set_visible(True) plt.gca().spines['right'].set_visible(True) plt.gca().spines['left'].set_visible(True) plt.gca().spines['bottom'].set_visible(True) #plt.rcParams['axes.linewidth'] = 2 #plt.rcParams['mathtext.fontset'] = 'cm' #plt.rcParams['mathtext.default'] = 'regular' #plt.rcParams['font.family'] = 'STIXGeneral' plt.savefig('Test1.jpg', dpi=300, bbox_inches='tight') plt.show() Continue reading...
