Foundation of Data Analytics (CDA-105)

(Unit-3)
Encoded Representation

An encoded representation, in the context of data analysis and machine learning, refers to a compact and
meaningful transformation of the original data into a different, often lower-dimensional, format. The goal of
encoding is to capture essential information from the data while reducing its dimensionality or complexity.
Encoded representations are particularly valuable in scenarios where the original data is high-dimensional or
noisy, and a more concise and informative representation is desired. Here are a few common types of encoded
representations:
Encoded Representation

Dimensionality Reduction:

One common form of encoding is

dimensionality reduction, where
the original features are
transformed into a lower-
dimensional space. Principal
Component Analysis (PCA) is a
classic example of dimensionality
reduction that creates an encoded
representation by projecting the
data onto a smaller set of principal
components, capturing the most
significant variation in the data.
Auto-Encoders

Autoencoders:

Autoencoders are a type of artificial neural network

designed for learning efficient representations of data.
They consist of an encoder network that compresses the
input data into a lower-dimensional representation
(encoding) and a decoder network that reconstructs the
input from this encoding. Autoencoders can learn
hierarchical and non-linear representations, making them
powerful for feature learning and dimensionality
reduction.
Sparse Coding

Sparse Coding:

•Sparse coding is a technique

where the representation of data
is sparse, meaning that only a
small subset of features is active.
This leads to a more compact and
informative encoding. Sparse
coding is commonly used in signal
processing and image analysis.
Encoder Representation

Feature Hashing:
Feature hashing is a technique that maps input
features to a fixed-size space using hash functions. It
is often used for encoding categorical features into a
high-dimensional space, providing a compromise
between dimensionality reduction and feature
representation.

A hash function is a mathematical function that takes an

input (or "message") and produces a fixed-size string of
characters, which is typically a hexadecimal or binary
representation. The output, commonly referred to as the
hash code or hash value, is a unique representation of the
input data. The primary purpose of a hash function is to
quickly and efficiently map data of arbitrary size to a fixed-
size value, often for the purpose of indexing, data
retrieval, or ensuring data integrity.
#- (hash function)

Deterministic:
A hash function is deterministic, meaning
that for a given input, it will always
produce the same hash value.
Fixed Output Size:
The output size of a hash function is fixed,
regardless of the size or length of the
input data. This fixed output size is a
defining characteristic of hash functions.
Efficient to Compute:
Hash functions are designed to be
computationally efficient, allowing for
quick computation of hash values even for
large amounts of data.
Spatial data Representation:

Spatial data representation involves the

methods and techniques used to model
and represent geographic information in
a digital format. Spatial data refers to
information that has a location
component, such as points on a map,
the boundaries of regions, or the
features of a landscape. Effectively
representing spatial data is crucial for
various applications, including
geographic information systems (GIS),
mapping, urban planning,
environmental analysis, and more. Here
are key concepts and methods in spatial
data representation:
Vector Data Model

Vector Data Model:

In the vector data model, spatial

features are represented as
points, lines, and polygons. Each
feature is described by its
geometry (coordinates) and
attributes (additional
information). Vector data is
suitable for representing
discrete features such as roads,
buildings, and administrative
boundaries.
Raster Data Model

Raster Data Model:

In the raster data model, spatial

information is represented as a grid
of cells or pixels, where each cell has
a value corresponding to a specific
attribute. Raster data is suitable for
continuous and regularly distributed
phenomena such as elevation
models, land cover, and satellite
imagery.
Co-ordinate System

Coordinate Systems:
•Coordinate systems define
how locations on the Earth's
surface are represented in
numerical terms. Common
coordinate systems include
latitude and longitude
(geographic coordinates) and
projected coordinate systems
(e.g., UTM, State Plane) that
provide a Cartesian coordinate
grid for specific regions.
Spatial Data Representation

Attributes and Tables:

In addition to spatial geometry, spatial data
often includes attribute data associated with
each spatial feature. Attribute tables store
non-spatial information about features,
allowing for the association of descriptive
information with specific locations.

Topology:
Topology defines the spatial relationships
and connectivity between geometric
features. It includes information about
adjacency, containment, and connectivity.
Topological relationships ensure data
integrity and consistency in spatial analysis.
Spatial Indexing

Spatial Indexing:

Spatial indexing structures,

such as quad-trees and R-
trees, are used to organize
spatial data efficiently. These
structures speed up spatial
queries and enable fast
retrieval of information
based on spatial
relationships.
Quadtree
3D Spatial data

Representing spatial
information in three
dimensions is essential for
applications such as urban
planning, geology, and
environmental modeling. 3D
spatial data includes
elevation models, 3D
buildings, and terrain models
Spatial data Representation: Cartography

Cartography is the discipline of creating

maps, which are visual representations of
the Earth's surface or other spatial
phenomena. Cartographers, professionals in
the field of cartography, use a combination
of science, art, and technology to design,
produce, and interpret maps. Maps serve
various purposes, including navigation,
communication of spatial information,
analysis of geographic patterns, and
representation of complex spatial
relationships. Here are key aspects of
cartography:
Cartography
Map Design:

Cartography involves the design and creation of maps, taking into consideration both aesthetic and functional
aspects. Map design includes decisions about color, symbols, typography, scale, and layout to effectively
convey information to the map reader.

Map Elements:
Maps typically include a variety of elements, such as:

• Title: Clearly indicates the subject or purpose of the map.

• Legend (or Key): Explains the symbols and colors used on the map.
• Scale: Represents the relationship between distances on the map and actual distances on the
Earth's surface.
• Compass Rose: Shows the orientation of the map (north, south, east, west).
• Insets: Additional small maps that provide more detail for specific areas.
• Grids and Coordinates: Help locate points on the map.
• Annotations: Text labels providing additional information.
GIS (Geographic Information System)

Modern cartography
often involves the use
of GIS technology,
which allows for the
creation, analysis, and
interpretation of
spatial data. GIS
integrates maps with
data, enabling more
sophisticated mapping
and spatial analysis.
Cartography

Digital Cartography:

With the advent of

computers, digital
cartography has become
prevalent. Cartographers
use Geographic
Information System (GIS)
software, computer-
aided design (CAD) tools,
and other software to
create and manipulate
digital maps.
Map Projection:

Map Projections:
•Representing the three-
dimensional Earth on a flat
surface involves distortion.
Cartographers use map
projections to transform the
Earth's curved surface into a
two-dimensional map.
Different map projections
are suited for different
purposes and areas.
Cartography

Cognitive Aspects:
Cartography also considers how people perceive and interpret maps. Cognitive
aspects involve understanding how map readers interpret symbols, colors, and
spatial relationships, which influences map design to enhance clarity and
effectiveness.

Historical Cartography:
Historical cartography involves the study of maps from the past, examining how
cartographic techniques and knowledge of the Earth's geography have evolved over
time. Historical maps provide insights into the perceptions and understanding of the
world in different eras.
Cartography

Cartographic Ethics:

Cartographers adhere to ethical considerations, ensuring that maps are accurate, unbiased, and
do not perpetuate stereotypes or misrepresent geographic information.
Cartography is a dynamic and interdisciplinary field that plays a crucial role in understanding and
representing the spatial aspects of our world. It continues to evolve with advancements in technology
and methods, contributing to the effective communication and analysis of spatial information.