Data Types and Structures

Data Types

• Numerical data is either discrete (integer) or continuous (float, double).
• Categorical data is either ordinal (ordered) or nominal (unordered).

Data Structures

Lists

• Lists are ordered, mutable, and can contain duplicate elements.
• Use my_list = [1, 2, 3] to create a list.
• Use my_list.append(4) to add a single element to the end.
• Use my_list.extend([5, 6]) to add multiple elements to the end.
• Use my_list.insert(0, 0) to insert an element at a specific index.
• Use my_list.remove(0) to remove the first occurrence of a specific element.
• Use my_list.pop(0) to remove and return the element at a specific index.
• Use del my_list to delete the item at index.
• Use my_list.clear() to remove all elements.

Dictionaries

• Dictionaries store key-value pairs and are mutable.
• Use my_dict = {'a': 1, 'b': 2} to create a dictionary.
• Use my_dict['c'] = 3 to add a new key-value pair.
• Use my_dict['a'] = 4 to modify the value of an existing key.
• Use del my_dict['a'] to delete a key-value pair.
• Use my_dict.pop('b') to remove and return an item.
• Use my_dict.clear() to remove key-value pairs from a dictionary.

Tuples

• Tuples are ordered and immutable.
• Use my_tuple = (1, 2, 3) to create a tuple.

Sets

• Sets are unordered and contain unique elements.
• Use my_set = {1, 2, 3} to create a set.
• Use my_set.add(4) to add an element.
• Use my_set.remove(4) to remove a given element.
• Use my_set.discard(4) to remove an element if it exists, without raising an error.
• Use my_set.pop() to remove an arbitrary element.
• Use my_set.clear() to remove all elements.

Data Preprocessing (with Pandas)

Handle Missing Data

• Use df.isnull() to identify missing values.
• Use df.notnull() to identify non-missing values.
• Use df.isnull().sum() to count missing values per column.
• Use df.dropna() to drop rows with any missing values.
• Use df.dropna(axis=1) to drop columns with any missing values.
• Use df.fillna(value) to fill missing values with a specified value.
• Use df.fillna(df.mean()) to fill missing values with the column mean.
• Use df.fillna(df.median()) to fill missing values with the column median.

Data Formatting

• Use df['column'].astype('int') to change the data type of a column.

Data Normalization

• Simple Scaling: df['column'] = df['column'] / df['column'].max()
• Min-Max: df['column'] = (df['column'] - df['column'].min()) / (df['column'].max() - df['column'].min())
• Z-score: df['column'] = (df['column'] - df['column'].mean()) / df['column'].std()

Data Discretization

• Use pd.cut() or pd.qcut() for data discretization.

Exploratory Data Analysis (EDA)

Descriptive Statistics

• Numerical data:
  • Use df.describe() to get descriptive statistics.
  • Use df['column'].mean() to get the mean.
  • Use df['column'].median() to get the median.
  • Use df['column'].mode() to get the mode.
  • Use df['column'].std() to get the standard deviation.
  • Use df['column'].var() to get the variance.
  • Use df['column'].min() to get the minimum value.
  • Use df['column'].max() to get the maximum value.
  • Use df['column'].sum() to get the sum.
  • Use df['column'].quantile(0.25) to get the 25th percentile.
  • Use df['column'].value_counts() to count occurrences.
• Categorical data:
  • Use df['column'].value_counts() to get frequency counts.

Grouping

• Use df.groupby('column')['another_column'].mean() to group by one column and get the mean of another column.

Visualization

• Histograms: plt.hist()
• Scatter plots: plt.scatter()
• Box plots: plt.boxplot()

Machine Learning

Model Evaluation Metrics

Regression

• Mean Absolute Error (MAE): $MAE = \frac{1}{n} \sum_{i=1}^{n} | y_i - \hat{y}_i |$
• Mean Squared Error (MSE): $MSE = \frac{1}{n} \sum_{i=1}^{n} (y_i - \hat{y}_i)^2$
• Root Mean Squared Error (RMSE): $RMSE = \sqrt{\frac{1}{n} \sum_{i=1}^{n} (y_i - \hat{y}_i)^2}$
• R-squared: $R^2 = 1 - \frac{\sum_{i=1}^{n} (y_i - \hat{y}i)^2}{\sum{i=1}^{n} (y_i - \bar{y})^2}$

Classification

• Accuracy: $\frac{TP + TN}{TP + TN + FP + FN}$
• Precision: $\frac{TP}{TP + FP}$
• Recall: $\frac{TP}{TP + FN}$
• F1-Score: $2 * \frac{Precision * Recall}{Precision + Recall}$
• AUC-ROC: Area under the Receiver Operating Characteristic curve

Common Algorithms

Regression

• Linear Regression
• Decision Tree
• Random Forest

Classification

• Logistic Regression
• Support Vector Machine (SVM)
• K-Nearest Neighbors (KNN)
• Decision Tree
• Random Forest

Clustering

• K-Means

Dimensionality Reduction

• Principal Component Analysis (PCA)

Ischemic Stroke Management Guidelines

Summary Points

• Rapid recognition and assessment of stroke symptoms are crucial.
• Immediate transfer to a stroke center is essential for eligible patients.
• Intravenous thrombolysis with alteplase should be considered within 4.5 hours of symptom onset.
• Endovascular thrombectomy is indicated for patients with large vessel occlusion within 24 hours of symptom onset.
• Blood pressure management, glucose control, and temperature regulation are important aspects of acute stroke care.
• Secondary prevention strategies should be implemented to reduce the risk of recurrent stroke.

Acute Management Algorithm

• Initial Assessment: Assess ABCs, obtain history, perform neurological examination (NIHSS), and rule out stroke mimics.
• Rapid Transport: Activate EMS and transport to the nearest stroke center, alerting the receiving hospital.
• Emergency Department Evaluation: Obtain CT or MRI brain imaging, perform lab tests, and assess eligibility for thrombolysis and/or thrombectomy.
• Thrombolysis: Administer intravenous alteplase within 4.5 hours of symptom onset if eligible, and monitor for bleeding complications.
• Endovascular Thrombectomy: Transfer to a thrombectomy-capable center if large vessel occlusion is present, and perform thrombectomy as soon as possible within 24 hours of symptom onset.
• Supportive Care: Manage blood pressure, glucose, and temperature, and prevent/treat complications.
• Secondary Prevention: Start antiplatelet therapy, initiate statin therapy, and address modifiable risk factors.

Key Considerations

• Blood Pressure Management:
  • Maintain blood pressure below 180/105 mmHg during acute phase.
  • Use labetalol, nicardipine, or other appropriate agents.
  • Avoid excessive blood pressure reduction, which can worsen ischemia.
• Glucose Control:
  • Maintain blood glucose between 140-180 mg/dL.
  • Use insulin for hyperglycemia.
  • Avoid hypoglycemia, which can mimic stroke symptoms.
• Temperature Regulation:
  • Treat fever aggressively with antipyretics.
  • Avoid hypothermia, which can worsen outcomes.

Discharge Planning

• Assess patient's functional status and rehabilitation needs.
• Provide education on secondary prevention strategies.
• Schedule follow-up appointments.

Performance Measures

• Percentage of patients receiving thrombolysis within 60 minutes of arrival
• Percentage of patients with large vessel occlusion undergoing thrombectomy within 24 hours of symptom onset
• Rate of symptomatic intracranial hemorrhage following thrombolysis

Matplotlib

What is Matplotlib?

• Matplotlib is a comprehensive Python library for creating static, animated, and interactive visualizations.
• It simplifies common tasks and tackles complex visualizations.
• Matplotlib is useful for:
  • Creating publication-quality plots
  • Making interactive figures
  • Customizing visual style and layout
  • Exporting to many file formats
  • Embedding in JupyterLab and graphical user interfaces
  • Utilizing third-party packages built on Matplotlib

Matplotlib Simple Plotting

• Import necessary libraries:
  • import matplotlib.pyplot as plt
  • import numpy as np
• Generate data:
  • t = np.arange(0.0, 2.0, 0.01)
  • s = 1 + np.sin(2 * np.pi * t)
• Plot data:
  • fig, ax = plt.subplots()
  • ax.plot(t, s)
• Customize plot:
  • ax.set(xlabel='time (s)', ylabel='voltage (mV)', title='About as simple as it gets, folks')
  • ax.grid()
• Save and display plot:
  • fig.savefig("test.png")
  • plt.show()