mapdata 3.8.0__tar.gz → 3.9.0__tar.gz

{mapdata-3.8.0/mapdata.egg-info → mapdata-3.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: mapdata
-Version: 3.8.0
+Version: 3.9.0
 Summary: An interactive map and table explorer for geographic coordinates in a spreadsheet, CSV file, or database
 Home-page: https://osdn.net/project/mapdata/
 Author: Dreas Nielsen
@@ -78,21 +78,28 @@ Selected rows in the data table can be exported to a CSV or spreadsheet file.
 *Mapdata* provides five different ways to show information about multiple data rows that all have the same geographic coordinates, and thus plot at the same location on the map.
 
 Data can also be displayed in several different types of plots: box plots, scatter
-plots, line charts, ECDF plots, Q-Q plots, Fisher-Jenks group plots, strip charts, 
-and counts of categorical and quantitative variables.  Plots
+plots, line charts, kernel density plots, ECDF plots, Q-Q plots, strip charts, 
+counts of categorical and quantitative variables, and others.  Plots
 can use either all data or only data values that are selected in the map and
 table.  Plots have a live connection to the data table and map, so when data selections are changed the plots are automatically updated.
 
 ![example plot](https://mapdata.readthedocs.io/en/latest/_images/UI_cat_stripchart.png)
 
-*Mapdata* can also display summaries of univariate and bivariate statistics for selected variables.  These summaries can use either all data in the data table or only the data that are highlighted on the map.  These summaries are updated immediately if different data are selected.
+*Mapdata* can also carry out some univariate, bivariate, and multivariate statistical
+summarizations and analyses.  Statistical analyses can use either all data in the data table
+or only the data that are highlighted on the map.  These summaries are updated immediately
+if different data are selected.
 
 ![Bivariate statistics](https://mapdata.readthedocs.io/en/latest/_images/Bivariate_dialog.png)
 
-In addition to univariate and bivariate statistics, and plots, *mapdata* can display the relationships between variables in the form of:
+Statistical summaries and analyses include:
+
+  * Bivariate ordinary least squares regression and trend evaluation.
 
   * A correlation matrix.
 
+  * Parametric and nonparametric analysis of variance (ANOVA).
+
   * A contingency table, using either categorical or numeric variables, with flexible
     specification of groups.  Tests of independence, risk ratio, odds ratio and related
     statistics, and conditional probabilities are all shown.

{mapdata-3.8.0 → mapdata-3.9.0}/README.md

@@ -36,21 +36,28 @@ Selected rows in the data table can be exported to a CSV or spreadsheet file.
 *Mapdata* provides five different ways to show information about multiple data rows that all have the same geographic coordinates, and thus plot at the same location on the map.
 
 Data can also be displayed in several different types of plots: box plots, scatter
-plots, line charts, ECDF plots, Q-Q plots, Fisher-Jenks group plots, strip charts, 
-and counts of categorical and quantitative variables.  Plots
+plots, line charts, kernel density plots, ECDF plots, Q-Q plots, strip charts, 
+counts of categorical and quantitative variables, and others.  Plots
 can use either all data or only data values that are selected in the map and
 table.  Plots have a live connection to the data table and map, so when data selections are changed the plots are automatically updated.
 
 ![example plot](https://mapdata.readthedocs.io/en/latest/_images/UI_cat_stripchart.png)
 
-*Mapdata* can also display summaries of univariate and bivariate statistics for selected variables.  These summaries can use either all data in the data table or only the data that are highlighted on the map.  These summaries are updated immediately if different data are selected.
+*Mapdata* can also carry out some univariate, bivariate, and multivariate statistical
+summarizations and analyses.  Statistical analyses can use either all data in the data table
+or only the data that are highlighted on the map.  These summaries are updated immediately
+if different data are selected.
 
 ![Bivariate statistics](https://mapdata.readthedocs.io/en/latest/_images/Bivariate_dialog.png)
 
-In addition to univariate and bivariate statistics, and plots, *mapdata* can display the relationships between variables in the form of:
+Statistical summaries and analyses include:
+
+  * Bivariate ordinary least squares regression and trend evaluation.
 
   * A correlation matrix.
 
+  * Parametric and nonparametric analysis of variance (ANOVA).
+
   * A contingency table, using either categorical or numeric variables, with flexible
     specification of groups.  Tests of independence, risk ratio, odds ratio and related
     statistics, and conditional probabilities are all shown.

{mapdata-3.8.0 → mapdata-3.9.0}/mapdata/mapdata.py

@@ -24,8 +24,8 @@
 #
 # ==================================================================
 
-version = "3.8.0"
-vdate = "2024-06-02"
+version = "3.9.0"
+vdate = "2024-06-13"
 
 copyright = "2023-2024"
 
@@ -86,7 +86,7 @@ import scipy.stats as spstats
 import scipy.sparse as sparse
 import statsmodels.api as sm
 from statsmodels.sandbox.stats.runs import runstest_1samp
-from statsmodels.stats.diagnostic import normal_ad, kstest_normal
+from statsmodels.stats.diagnostic import normal_ad, lilliefors, kstest_normal
 try:
 	import pymannkendall as mk
 	mk_available = True
@@ -1380,6 +1380,32 @@ def check_rowtable(rowtable, headers, parent, tablename=None):
 			msg += f"Row {rowno+1} has {len(rowtable[rowno])} columns but should have {n_headers}.\nBad row: {rowtable[rowno]}"
 			fatal_error(msg, kwargs={'parent': parent})
 
+def clean_missing(dataset, column_indexes):
+	# Removes rows for a table as a list of columns if there is a missing value in any
+	# of the columns specified by the indexes
+	clean_data = [[] for _ in dataset]
+	for i in range(len(dataset[0])):
+		ok = True
+		for col in column_indexes:
+			if dataset[col][i] is None or dataset[col][i] == '':
+				ok = False
+		if ok:
+			for col in column_indexes:
+				clean_data[col].append(dataset[col][i])
+	return clean_data
+
+
+def spread_by_groups(grouplist, valuelist):
+	# Split 'valuelist' into several lists, one for each group in grouplist.
+	# The lists must have the same length.
+	# Returns a tuple of group names (as strings) and a list of lists.  Missing values are eliminated.
+	grpstrs = [str(g) for g in grouplist]
+	grp_vals = sorted(list(set(grpstrs)))
+	ds = list(zip(grpstrs, valuelist))
+	dspread = []
+	for g in grp_vals:
+		dspread.append([d[1] for d in ds if d[0] == g and d[1] is not None])
+	return grp_vals, dspread
 
 def subset_by_groups(dataset, grouplist):
 	# Returns a dictionary with keys of unique elements of 'grouplist' and values
@@ -1975,6 +2001,8 @@ class MapUI(object):
 		self.univar_list = []
 		# List of BivarStatsDialog objects, so they can be updated.
 		self.bivar_list = []
+		# List of ANOVADialog objects, so they can be updated.
+		self.anova_list = []
 		# List of FitDistDialog objects, so they can be updated.
 		self.fitdist_list = []
 		# List of TSNEDialog objects, so they can be updated.
@@ -2110,7 +2138,7 @@ class MapUI(object):
 				self.cancel()
 		else:
 			if args.lat is None or args.lon is None:
-				warning("Latitude and longitude columns must be specified on the command line with other data source arguments.", kwargs={'parent': self.win})
+				warning("Latitude and longitude columns must be specified with other data source arguments.", kwargs={'parent': self.win})
 				sdsd = SelDataSrcDialog(parent=self.win, mapui=self)
 				src_name, label_col, lat_col, lon_col, crs, symbol_col, color_col, message, headers, rows = sdsd.select()
 				if src_name is None:
@@ -2829,6 +2857,9 @@ class MapUI(object):
 		for dlg in self.bivar_list:
 			if unconditional or dlg.sel_only_var.get() == "1":
 				dlg.q_recalc()
+		for dlg in self.anova_list:
+			if unconditional or dlg.sel_only_var.get() == "1":
+				dlg.q_recalc()
 		for dlg in self.fitdist_list:
 			if unconditional or dlg.sel_only_var.get() == "1":
 				dlg.q_recalc()
@@ -2967,9 +2998,14 @@ class MapUI(object):
 			self.bivar_list.remove(bivar_dlg)
 		except:
 			pass
+	def remove_anova(self, anova_dlg):
+		try:
+			self.anova_list.remove(anova_dlg)
+		except:
+			pass
 	def remove_fitdist(self, fitdist_dlg):
 		try:
-			self.fitdist_list.remove(bivar_dlg)
+			self.fitdist_list.remove(fitdist_dlg)
 		except:
 			pass
 	def remove_corrmat(self, corrmat_dlg):
@@ -3534,6 +3570,12 @@ class MapUI(object):
 			self.bivar_list.append(dlg)
 			dlg.show()
 
+		def anova_stats():
+			self.wait_for_data_types()
+			dlg = ANOVADialog(self, self.data_types)
+			self.anova_list.append(dlg)
+			dlg.show()
+
 		def fitdist_stats():
 			self.wait_for_data_types()
 			dlg = FitDistDialog(self, self.data_types)
@@ -3660,6 +3702,7 @@ class MapUI(object):
 		stats_menu.add_command(label="Fit univariate distribution", command = fitdist_stats, underline=0)
 		stats_menu.add_command(label="Bivariate statistics", command = bivar_stats, underline=0)
 		stats_menu.add_command(label="Correlation matrix", command = corr_matrix, underline=0)
+		stats_menu.add_command(label="One-way ANOVA", command = anova_stats, underline=8)
 		stats_menu.add_command(label="Contingency table", command = cont_table, underline=2)
 		stats_menu.add_command(label="ROC curve", command = roc_curve, underline=0)
 		stats_menu.add_command(label="t-SNE analysis", command = tsne_anal, underline=0)
@@ -5886,18 +5929,8 @@ class PlotDialog(object):
 				self.plot_data = [x_vals, x_counts]
 				self.plot_data_labels = [self.data_labels[0], "Count"]
 			elif plot_type in ("Box plot", "Histogram", "Kernel density (KD) plot", "Stripchart", "Violin plot"):
-				# The groupby variable may or may not be set
 				if self.groupby_var.get() != '':
-					# Ensure that all values of the labeling variable are strings
-					self.dataset[1] = [str(lbl) for lbl in self.dataset[1]]
-					grp_vals = list(set(self.dataset[1]))
-					grp_vals.sort()
-					ds = list(zip(self.dataset[1], self.dataset[0]))
-					plot_data = []
-					for g in grp_vals:
-						plot_data.append([d[1] for d in ds if d[0] == g])
-					self.plot_data = plot_data
-					self.plot_data_labels = grp_vals
+					self.plot_data_labels, self.plot_data = spread_by_groups(self.dataset[1], self.dataset[0])
 				else:
 					self.plot_data = [self.dataset[0]]
 					self.plot_data_labels = [self.data_labels[0]]
@@ -7295,10 +7328,10 @@ class UnivarStatsDialog(object):
 		self.categ_columns.sort()
 		self.dnames = ["Variable", " N ", "Min.", "Max.", "Mean", "Median", "Mode", \
 				"Geo. mean", "Std.Dev.", "C.V.", "Sum", "5th %ile", "95th %ile", "Anderson-Darling p", \
-				"Lillefors p", "Rosner's outliers", "Tukey outliers"]
+				"Lilliefors p", "Omnibus normality p", "Rosner's outliers", "Tukey outliers"]
 		self.logdnames = ["Variable", " N ", "Min.", "Max.", "Mean", "Median", "Mode", \
-				"Std.Dev.", "C.V.", "Sum", "5th %ile", "95th %ile", "Anderson-Darling p", "Lillefors p", \
-				"Rosner's outliers", "Tukey outliers"]
+				"Std.Dev.", "C.V.", "Sum", "5th %ile", "95th %ile", "Anderson-Darling p", "Lilliefors p", \
+				"Omnibus normality p", "Rosner's outliers", "Tukey outliers"]
 		self.statdata = []
 		self.dlg.bind("<Control-s>")
 		self.dlg.bind("<Control-z>")
@@ -7475,21 +7508,32 @@ class UnivarStatsDialog(object):
 				da = np.array(d)
 				normstats = []
 				if stdev_d != 0.0:
+					# Anderson-Darling
 					try:
 						adval, adpval = normal_ad(da)
 						normstats.append("%.2E" % adpval)
 					except:
 						normstats.append("NC")
+					# Lilliefors
 					try:
-						lfval, lfpval = kstest_normal(da)
+						lfval, lfpval = lilliefors(da)
 						normstats.append("%.2E" % lfpval)
 					except:
 						normstats.append("NC")
+					# Omnibus
+					if len(da) > 19:
+						try:
+							stat, p = spstats.normaltest(da)
+							normstats.append("%.2E" % p)
+						except:
+							normstats.append("NC")
+					else:
+						normstats.append("NC")
 				else:
-					normstats = ["NC", "NC"]
+					normstats = ["NC", "NC", "NC"]
 				dd.extend(normstats)
 			else:
-				dd.extend([None, None])
+				dd.extend(["NC", "NC", "NC"])
 			#=== Outlier evaluation
 			if len(d) > 14:
 				if stdev_d > 0.0:
@@ -7558,6 +7602,7 @@ class UnivarStatsDialog(object):
 							self.logstatdata.append([grp, self.data_labels[var]] + logdd)
 				dnames = ["Group"] + self.dnames
 				logdnames = ["Group"] + self.logdnames
+		breakpoint()
 		tframe, tdata = treeview_table(self.data_page, self.statdata, dnames)
 		tframe.grid(row=0, column=0, sticky=tk.NSEW)
 		ltframe, ltdata = treeview_table(self.log_page, self.logstatdata, logdnames)
@@ -7721,6 +7766,17 @@ class BivarStatsDialog(object):
 		self.residfig_canvas.get_tk_widget().pack(side=tk.TOP, fill=tk.BOTH, expand=1)
 		#self.resid_nav.update()
 		plot_pages.add(resid_frame, text="Residuals")
+		# Şen 2012 trend plot
+		sen_frame = tk.Frame(plot_pages)
+		sen_frame.rowconfigure(0, weight=1)
+		sen_frame.columnconfigure(0, weight=1)
+		sen_frame.pack(side=tk.TOP, fill=tk.BOTH, expand=1)
+		self.senfig = Figure(dpi=100)
+		self.senfig.set_figheight(2)
+		self.senfig.set_figwidth(2)
+		self.senfig_canvas = FigureCanvasTkAgg(self.senfig, sen_frame)
+		self.senfig_canvas.get_tk_widget().pack(side=tk.TOP, fill=tk.BOTH, expand=1)
+		plot_pages.add(sen_frame, text="Şen trend")
 
 		# Initialize output frames
 		self.clear_output()
@@ -7842,6 +7898,9 @@ class BivarStatsDialog(object):
 		self.residfig.clear()
 		self.resid_axes = self.residfig.add_subplot(111)
 		self.residfig_canvas.draw()
+		self.senfig.clear()
+		self.sen_axes = self.senfig.add_subplot(111)
+		self.senfig_canvas.draw()
 
 	def show_data(self, *args):
 		if self.dataset is not None:
@@ -7995,6 +8054,24 @@ class BivarStatsDialog(object):
 				self.plot_axes.scatter(self.dataset[0], self.dataset[1], alpha=self.alpha)
 				self.plotfig_canvas.draw()
 				self.residfig_canvas.draw()
+				# Plot of Şen trend
+				ylen = len(self.dataset[1])
+				if ylen > 3:
+					if ylen % 2:
+						ysplit = int((ylen - 1)/2)
+						firsthalf = self.dataset[1][1:][:ysplit]
+						secondhalf = self.dataset[1][1:][ysplit:]
+					else:
+						ysplit = int(ylen/2)
+						firsthalf = self.dataset[1][:ysplit]
+						secondhalf = self.dataset[1][ysplit:]
+					self.sen_axes.scatter(firsthalf, secondhalf)
+					mid = statistics.median(self.dataset[1])
+					self.sen_axes.axline((mid,mid), slope=1, color="grey", alpha=0.65)
+					self.sen_axes.set_xlabel("First half of "+self.data_labels[1])
+					self.sen_axes.set_ylabel("Second half of "+self.data_labels[1])
+					self.senfig_canvas.draw()
+				#
 				self.plot_nav.update()
 				self.dlg.bind("<Alt-a>", self.set_alpha)
 				self.dlg.bind("<Alt-t>", self.set_title)
@@ -8289,6 +8366,296 @@ class CorrMatrixDialog(object):
 			pass
 
 
+class ANOVADialog(object):
+	def __init__(self, parent, column_specs):
+		self.parent = parent
+		self.column_specs = column_specs
+		self.alpha = 0.45
+		self.theilsen = False
+		self.plot_title = None
+		self.dlg = tk.Toplevel()
+		self.dlg.title("Analysis of Variance (ANOVA)")
+		self.dlg.columnconfigure(0, weight=1)
+		# Data
+		self.dataset = None
+		self.data_labels = None
+		self.group_data = None
+		self.group_labels = None
+		self.numeric_columns = sorted([c[0] for c in self.column_specs if c[1] in ("int", "float")])
+		self.categ_columns = sorted([c[0] for c in self.column_specs if c[1] == "string"])
+		self.output_colhdrs = ["Test", "Statistic", "Value", "p value"]
+		self.dist_colhdrs = ["Group", "N", "Mean", "Median", "Variance", "Skewness", "Kurtosis", "Anderson-Darling p", "Lilliefors p", "Omnibus normality p"]
+		self.statdata = []
+		# Message
+		prompt_frame = tk.Frame(self.dlg, borderwidth=5)
+		prompt_frame.grid(row=0, column=0, sticky=tk.N+tk.EW, pady=(3,3))
+		prompt_frame.columnconfigure(0, weight=1)
+		msg_lbl = ttk.Label(prompt_frame, width=70, text="Select numeric and grouping variables")
+		msg_lbl.grid(row=0, column=0, sticky=tk.W, padx=(6,6), pady=(3,3))
+		def wrap_msg(event):
+			msg_lbl.configure(wraplength=event.width - 5)
+		msg_lbl.bind("<Configure>", wrap_msg)
+
+		# Controls
+		ctrl_frame = tk.Frame(self.dlg, borderwidth=5)
+		ctrl_frame.grid(row=1, column=0, sticky=tk.N+tk.EW)
+
+		self.x_var = tk.StringVar(ctrl_frame, "")
+		x_lbl = ttk.Label(ctrl_frame, text="X column:")
+		x_lbl.grid(row=0, column=0, sticky=tk.E, padx=(6,3), pady=(3,3))
+		self.x_sel = ttk.Combobox(ctrl_frame, state=tk.NORMAL, textvariable=self.x_var, values=self.numeric_columns, width=24)
+		self.x_sel.grid(row=0, column=1, sticky=tk.W, padx=(3,6), pady=(3,3))
+		self.x_sel.bind("<<ComboboxSelected>>", self.q_recalc)
+
+		self.xlog_var = tk.StringVar(ctrl_frame, "0")
+		self.xlog_ck = ttk.Checkbutton(ctrl_frame, text="Log X", state=tk.NORMAL, command=self.q_recalc, variable=self.xlog_var,
+				onvalue="1", offvalue="0")
+		self.xlog_ck.grid(row=0, column=2, sticky=tk.W, padx=(6,6), pady=(3,3))
+
+		self.groupby_var = tk.StringVar(ctrl_frame, "")
+		groupby_lbl = ttk.Label(ctrl_frame, text="Group by:")
+		groupby_lbl.grid(row=1, column=0, sticky=tk.E, padx=(6,3), pady=(3,3))
+		self.groupby_sel = ttk.Combobox(ctrl_frame, state=tk.NORMAL, textvariable=self.groupby_var, values=self.categ_columns, width=24)
+		self.groupby_sel.grid(row=1, column=1, sticky=tk.W, padx=(3,6), pady=(3,3))
+		self.groupby_sel.bind("<<ComboboxSelected>>", self.q_recalc)
+
+		self.sel_only_var = tk.StringVar(ctrl_frame, "0")
+		self.sel_only_ck = ttk.Checkbutton(ctrl_frame, text="Selected data only", command=self.q_recalc, variable=self.sel_only_var,
+				onvalue="1", offvalue="0")
+		self.sel_only_ck.grid(row=2, column=0, columnspan=2, sticky=tk.W, padx=(6,3), pady=(3,3))
+
+		# Frames for output tables
+		self.content_frame = tk.Frame(self.dlg)
+		self.content_frame.grid(row=2, column=0, sticky=tk.NSEW)
+		self.dlg.rowconfigure(2, weight=1)
+		self.dlg.columnconfigure(0, weight=1)
+		self.content_frame.rowconfigure(0, weight=1)
+		self.content_frame.columnconfigure(0, weight=1)
+		# with tabbed Notebook for two different tables
+		table_pages = ttk.Notebook(self.content_frame)
+		table_pages.grid(row=0, column=0, sticky=tk.NSEW)
+		table_pages.rowconfigure(0, weight=1)
+		table_pages.columnconfigure(0, weight=1)
+		# Frame for ANOVA statistics
+		self.stat_frame = tk.Frame(table_pages)
+		self.stat_frame.rowconfigure(0, weight=1)
+		self.stat_frame.columnconfigure(0, weight=1)
+		table_pages.add(self.stat_frame, text="ANOVA")
+		# Frame for subgroup distribution statistics
+		self.dist_frame = tk.Frame(table_pages)
+		self.dist_frame.rowconfigure(0, weight=1)
+		self.dist_frame.columnconfigure(0, weight=1)
+		table_pages.add(self.dist_frame, text="Distributions")
+
+		# Initialize output frames
+		self.clear_output()
+
+		# Buttons
+		btn_frame = tk.Frame(self.dlg, borderwidth=3, relief=tk.RIDGE)
+		btn_frame.columnconfigure(0, weight=1)
+		btn_frame.grid(row=3, column=0, sticky=tk.EW, pady=(3,3))
+		btn_frame.columnconfigure(0, weight=0)
+		btn_frame.columnconfigure(1, weight=0)
+		btn_frame.columnconfigure(2, weight=1)
+		self.canceled = False
+		self.help_btn = ttk.Button(btn_frame, text="Help", command=self.do_help, underline=0)
+		self.help_btn.grid(row=0, column=0, sticky=tk.W, padx=(6,3))
+		self.dlg.bind("<Alt-h>", self.do_help)
+		self.data_btn = ttk.Button(btn_frame, text="Source Data", state=tk.DISABLED, command=self.show_data, underline=0)
+		self.data_btn.grid(row=0, column=1, sticky=tk.W, padx=(3,3))
+		self.group_btn = ttk.Button(btn_frame, text="Group Data", state=tk.DISABLED, command=self.show_group_data, underline=0)
+		self.group_btn.grid(row=0, column=2, sticky=tk.W, padx=(3,3))
+		close_btn = ttk.Button(btn_frame, text="Close", command=self.do_close, underline=0)
+		close_btn.grid(row=0, column=3, sticky=tk.E, padx=(6,6))
+		self.dlg.bind("<Alt-c>", self.do_close)
+		self.dlg.bind("<Escape>", self.do_close)
+		center_window(self.dlg)
+		raise_window(self.dlg)
+
+	def do_help(self, *args):
+		webbrowser.open("https://mapdata.readthedocs.io/en/latest/dialogs.html#anova-dialog", new=2, autoraise=True)
+
+	def get_data(self):
+		self.data_btn["state"] = tk.DISABLED
+		self.dataset = None
+		self.data_labels = None
+		self.group_data = None
+		self.group_labels = None
+		self.dlg.bind("<Alt-s>")
+		self.dlg.bind("<Alt-g>")
+		column_list = [self.x_var.get(), self.groupby_var.get()]
+		# Get either only the selected data or all data.
+		if self.sel_only_var.get() == "1":
+			dataset = self.parent.get_sel_data(column_list)
+		else:
+			dataset = self.parent.get_all_data(column_list)
+		if dataset is None or len(dataset[0]) == 0:
+			self.data_btn["state"] = tk.DISABLED
+			self.group_btn["state"] = tk.DISABLED
+		else:
+			# Drop missing data
+			dataset = clean_missing(dataset, [0,1])
+			# Convert nominally numeric values to floats
+			float_vals = [conv_float(v) for v in dataset[0]]
+			# Log-transform data if specified.
+			if self.xlog_ck["state"] != tk.DISABLED and self.xlog_var.get() == "1":
+				log_error = False
+				log_data = []
+				for i in range(len(float_vals)):
+					try:
+						log_data.append(math.log10(float_vals[i]))
+					except:
+						log_error = True
+						self.xlog_var.set("0")
+						self.xlog_ck["state"] = tk.DISABLED
+						break
+				if log_error:
+					warning("Data cannot be log-transformed", {'parent':self.dlg})
+				else:
+					float_vals = log_data
+			self.group_labels, self.group_data = spread_by_groups(dataset[1], float_vals)
+			self.group_btn["state"] = tk.NORMAL
+			self.dataset = [float_vals, dataset[1]]
+			if self.xlog_ck["state"] != tk.DISABLED and self.xlog_var.get() == "1":
+				self.data_labels = ["Log10 of %s" % column_list[0]]
+			else:
+				self.data_labels = [column_list[0]]
+			self.data_labels.append(column_list[1])
+			self.data_btn["state"] = tk.NORMAL
+			self.dlg.bind("<Alt-s>", self.show_data)
+			self.dlg.bind("<Alt-g>", self.show_group_data)
+
+	def clear_output(self):
+		for ctl in self.stat_frame.winfo_children():
+			ctl.destroy()
+		for ctl in self.dist_frame.winfo_children():
+			ctl.destroy()
+		self.dlg.bind("<Control-s>")
+		self.dlg.bind("<Control-z>")
+		tframe, tdata = treeview_table(self.stat_frame, [], self.output_colhdrs)
+		tframe.grid(row=0, column=0, sticky=tk.NSEW)
+		tframe, tdata = treeview_table(self.dist_frame, [], self.dist_colhdrs)
+		tframe.grid(row=0, column=0, sticky=tk.NSEW)
+		self.dlg.bind("<Alt-s>")
+		self.dlg.bind("<Alt-g>")
+
+	def show_data(self, *args):
+		if self.dataset is not None:
+			show_columnar_table(self.dlg, "Source Data", "Selected data:", self.dataset, self.data_labels[0:len(self.dataset)], \
+					"Source Data for ANOVA")
+
+	def show_group_data(self, *args):
+		if self.group_data is not None:
+			show_columnar_table(self.dlg, "Grouped Data", "Data for each group:", self.group_data, self.group_labels, "Data for ANOVA")
+
+	def q_recalc(self, get_data=True, *args):
+		if self.x_var.get() != '' and self.groupby_var.get() != '':
+			if get_data or self.dataset is None:
+				self.get_data()
+			if self.group_data is not None and len(self.group_data) > 1 and max([len(d) for d in self.group_data]) > 1:
+				self.recalc()
+			else:
+				self.clear_output()
+
+	def recalc(self):
+		self.clear_output()
+		self.statdata = []
+		self.distdata = []
+		if self.group_data is not None:
+			group_lengths = [len(g) for g in self.group_data]
+			# One-way ANOVA
+			F, p = spstats.f_oneway(*self.group_data)
+			self.statdata.append(["One-way ANOVA", "F", fp_display(F), fp_display(p)])
+			# Kruskal-Wallace test
+			H, p = spstats.kruskal(*self.group_data)
+			self.statdata.append(["Kruskal-Wallis", "H", fp_display(H), fp_display(p)])
+			# Alexander-Govern test
+			if min(group_lengths) > 1:
+				ag = spstats.alexandergovern(*self.group_data)
+				A = ag.statistic
+				p = ag.pvalue
+				self.statdata.append(["Alexander-Govern", "A", fp_display(A), fp_display(p)])
+			# Levene's test
+			W, p = spstats.levene(*self.group_data, center="mean")
+			self.statdata.append(["Levene", "W", fp_display(W), fp_display(p)])
+			# Brown-Forsythe test
+			F, p = spstats.levene(*self.group_data, center="median")
+			self.statdata.append(["Brown-Forsythe", "F", fp_display(F), fp_display(p)])
+			# Bartlett's test
+			try:
+				X2, p = spstats.bartlett(*self.group_data)
+			except:
+				pass
+			else:
+				if not math.isnan(X2) and not math.isnan(p):
+					self.statdata.append(["Bartlett", "X2", fp_display(X2), fp_display(p)])
+			# Distribution statistics for each group
+			for i in range(len(self.group_labels)):
+				# Group, N, Mean, Median, Variance, Skewness, Kurtosis, Normality p]
+				lbl = self.group_labels[i]
+				d = self.group_data[i]
+				dd = [lbl, group_lengths[i], fp_display(statistics.fmean(d))]
+				dd.append(fp_display(statistics.median(d)))
+				var = None
+				if group_lengths[i] > 1:
+					try:
+						var = statistics.variance(d)
+						dd.append(fp_display(var))
+					except:
+						dd.append("NC")
+				else:
+					dd.append("NC")
+				dd.extend([fp_display(spstats.skew(d)), fp_display(spstats.kurtosis(d))])
+				# Normality tests
+				if var is not None:
+					try:
+						adval, adpval = normal_ad(da)
+						dd.append("%.2E" % adpval)
+					except:
+						dd.append("NC")
+					try:
+						lfval, lfpval = lilliefors(da, dist='norm')
+						dd.append("%.2E" % lfpval)
+					except:
+						dd.append("NC")
+					if len(self.group_data[i]) > 19:
+						try:
+							stat, p = spstats.normaltest(d)
+						except:
+							dd.append("NC")
+						else:
+							dd.append("%.2E" % p)
+					else:
+						dd.append("NC")
+				else:
+					dd.extend(["NC", "NC", "NC"])
+
+				self.distdata.append(dd)
+		if len(self.statdata) > 0:
+			tframe, tdata = treeview_table(self.stat_frame, self.statdata, self.output_colhdrs)
+			tframe.grid(row=0, column=0, stick=tk.NSEW)
+			tframe, tdata = treeview_table(self.dist_frame, self.distdata, self.dist_colhdrs)
+			tframe.grid(row=0, column=0, stick=tk.NSEW)
+			self.dlg.bind("<Control-s>", self.save_table)
+			self.dlg.minsize(width=700, height=400)
+			self.dlg.bind("<Control-s>", self.save_table)
+			self.dlg.bind("<Control-z>", self.save_norm_table)
+
+	def save_table(self, *args):
+		export_data_table(self.output_colhdrs, self.statdata, sheetname="ANOVA statistsics")
+	def save_norm_table(self, *args):
+		export_data_table(self.dist_colhdrs, self.distdata, sheetname="Normality statistsics")
+	def show(self):
+		self.dlg.wait_window(self.dlg)
+	def do_close(self, *args):
+		self.parent.remove_anova(self)
+		try:
+			self.dlg.destroy()
+		except:
+			pass
+
+
+
 class TSNEDialog(object):
 	def __init__(self, parent, column_specs, prohibited_columns):
 		self.parent = parent
@@ -16366,8 +16733,8 @@ def read_all_config(argobj):
 	
 	# Update 'argobj'
 	for att in ("db_type", "server", "port", "database", "user", "table", "script", \
-			"lon", "lat", "id", "projection", "message", "file", "sheet", "imagefile", \
-			"imagewait"):
+			"lon", "lat", "id", "projection", "message", "file", "sheet", "symbol", \
+			"color", "message", "imagefile", "imagewait"):
 		if hasattr(cl_arg_defaults, att) and getattr(argobj, att) is None:
 			setattr(argobj, att, getattr(cl_arg_defaults, att))
 
@@ -16696,7 +17063,7 @@ def clparser():
 			help="The name of the column in the data file containing location identifiers or labels.")
 	parser.add_argument("-k", "--db_type", type=str, choices=['p', 's', 'l', 'm', 'k', 'o', 'f'], dest="db_type", default=None,
 			help="Database type: 'p'-PostgreSQL; 's'-SQL Server; 'l'-SQLite, 'm'-MySQL/MariaDB, 'k'-DuckDB, 'o'-Oracle, 'f'-Firebird.")
-	parser.add_argument('-m', '--message', dest='message', default='Map display',
+	parser.add_argument('-m', '--message', dest='message',
 			help='A message to display above the map.')
 	parser.add_argument("-n", "--no_passwd", action="store_true", dest="no_passwd", default=False,
 			help="Do not prompt for the password when the user is specified (default is to prompt).")

{mapdata-3.8.0 → mapdata-3.9.0/mapdata.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: mapdata
-Version: 3.8.0
+Version: 3.9.0
 Summary: An interactive map and table explorer for geographic coordinates in a spreadsheet, CSV file, or database
 Home-page: https://osdn.net/project/mapdata/
 Author: Dreas Nielsen
@@ -78,21 +78,28 @@ Selected rows in the data table can be exported to a CSV or spreadsheet file.
 *Mapdata* provides five different ways to show information about multiple data rows that all have the same geographic coordinates, and thus plot at the same location on the map.
 
 Data can also be displayed in several different types of plots: box plots, scatter
-plots, line charts, ECDF plots, Q-Q plots, Fisher-Jenks group plots, strip charts, 
-and counts of categorical and quantitative variables.  Plots
+plots, line charts, kernel density plots, ECDF plots, Q-Q plots, strip charts, 
+counts of categorical and quantitative variables, and others.  Plots
 can use either all data or only data values that are selected in the map and
 table.  Plots have a live connection to the data table and map, so when data selections are changed the plots are automatically updated.
 
 ![example plot](https://mapdata.readthedocs.io/en/latest/_images/UI_cat_stripchart.png)
 
-*Mapdata* can also display summaries of univariate and bivariate statistics for selected variables.  These summaries can use either all data in the data table or only the data that are highlighted on the map.  These summaries are updated immediately if different data are selected.
+*Mapdata* can also carry out some univariate, bivariate, and multivariate statistical
+summarizations and analyses.  Statistical analyses can use either all data in the data table
+or only the data that are highlighted on the map.  These summaries are updated immediately
+if different data are selected.
 
 ![Bivariate statistics](https://mapdata.readthedocs.io/en/latest/_images/Bivariate_dialog.png)
 
-In addition to univariate and bivariate statistics, and plots, *mapdata* can display the relationships between variables in the form of:
+Statistical summaries and analyses include:
+
+  * Bivariate ordinary least squares regression and trend evaluation.
 
   * A correlation matrix.
 
+  * Parametric and nonparametric analysis of variance (ANOVA).
+
   * A contingency table, using either categorical or numeric variables, with flexible
     specification of groups.  Tests of independence, risk ratio, odds ratio and related
     statistics, and conditional probabilities are all shown.