1.

Big Data
- Refers to da ta s ets whos e s ize is beyond the a bility of typica l da ta ba s e
s oftwa re tools to ca pture, s tore, ma na ge, a nd a na lyze.
- Size from tera bytes (10 12 bytes ), or peta bytes (10 15 bytes ), to exa bytes (10 18
bytes ).
- The notion of wha t is Big da ta will depend on the indus try, how da ta is us ed,
how much his torica l da ta is involved a nd ma ny other cha ra cteris tics .
- Cha ra cteris tics :
- Volum e
- Velocity
- Va riety
- Vera city

3
1. Big Data
- Cha ra cteris tics :
- Volume: The enorm ous s ize of da ta .
- Da ta tha t is a utom a tica lly genera ted tends to ha ve big a m ount.
- Exa m ples : Sens or da ta ; da ta from s ca nning equipm ent; or da ta from m ea s urem ent
devices .
- Velocity: The s peed a t which da ta is crea ted, a ccum ula ted, inges ted, a nd proces s ed.
- Variety: Big Da ta ca n be s tructured, uns tructured, a nd s em i-s tructured. Collected from
different s ources
- Structured data: In Structured s chem a , a long with a ll the required colum ns . It is in a
ta bula r form . Structured Da ta is s tored in the rela tiona l da ta ba s e m a na gem ent s ys tem .
- Semi-structured: In Sem i-s tructured, the s chem a is not a ppropria tely defined, e.g.,
J SON, XML, CSV, TSV, a nd em a il.
- Unstructured Data: All the uns tructured files , log files , a udio files , a nd im a ge files a re
included in the uns tructured da ta .
- Veracity: Refers to the a ccura cy of your da ta . The credibility of the s ource, a nd the s uita bility
of da ta for its ta rget a udience.

4
1. Big Data
- There is a big ra nge of big da ta a pplica tions a cros s mos t indus tries :
Tra ns porta tion, educa tion, hea lthca re, a dvertis ing a nd ma rketing, ba nking a nd
fina ncia l s ervices . Some key orga niza tiona l functions fueled by big da ta :
- Productivity and cost optimization: Ma nufa cturers utilize big da ta to im prove qua lity a nd
output while m inim izing cos t. Reta ilers ca n a ls o us e big da ta to revea l hidden cha nges in
cons um er beha vior.

- Product development: Com pa nies a na lyze a nd m odel a ra nge of big da ta inputs to foreca s t
cus tom er dem a nd a nd m a ke predictions to wha t kinds of new products a nd product fea tures
tha t m os t va lua ble to us ers .

- Smart technologies: Big da ta pla ys a crucia l role in collecting a nd a na lyzing da ta from

s ens ors , ca m era s , a nd IoT devices us ed every da y.

- Cybersecurity: Big da ta ca n be us ed for rea l-tim e threa t detection, log a na lys is , a nd network
m onitoring.
5
2. Relational Databas es (RDBMS)
- Overview of rela tiona l da ta ba s es :
- Structured Data Model: Data is organized into tables , where each table has a
unique nam e and cons is ts of rows and colum ns .

- Tables and Relationships: In relational databas es , data is s tored in m ultiple

tables , and relations hips between tables are es tablis hed us ing keys .

- Data Integrity Constraints: Relational databas es enforce data integrity cons traints
to m aintain data cons is tency and accuracy.

6
2. Relational Databas es (RDBMS)
- Cha llenges of Rela tiona l Da ta ba s es with Big Da ta
- Scalability: Not des igned to handle the m as s ive volum e of data. It becom es
expens ive and technically challenging to s cale to handle m as s ive datas ets .

- Schema inflexibility: Relationa l databas es are des igned for s tructured data with
predefined s chem as , and they can s truggle to handle the volum e, velocity, and
variety of data.

- Performance: J oin operations can becom e s low and res ource-cons um ing as the
data volum e increas es . Additionally, relationa l databas es m ay not be optim ized
for the dis tributed com puting architectures

7
 3. MongoDB
MongoDB is a popula r, open-s ource NoSQL (Not Only SQL) da ta ba s e ma na gement
s ys tem known for its flexibility, s ca la bility, a nd ea s e of us e. It fa lls under the ca tegory
of document-oriented da ta ba s es , where da ta is s tored in flexible, J SON-like documents
ins tea d of tra ditiona l row-a nd-column ta bles us ed in rela tiona l da ta ba s es .
- Horizontal scalability: In contras t to RDBMS, which are s caled vertically. MongoDB is
des igned to s cale horizontally acros s m ultiple nodes . Es s ential for s erving the m as s ive
am ounts of data
- Schemaless: Unlike relational databas es that enforce a rigid s chem a, MongoDB's
docum ent-oriented data m odel allows for the s torage of divers e and uns tructured data
types com m only found in big data applications .
- Real-time Analytics: MongoDB's fram ework enables real-tim e analytics on large
datas ets by s upporting com plex operations directly within the
databas e.
8
The advantages of NoSQL databases
over relational databases

9
Database Architecture
Typical Data Architecture

The era of one-size-fits-all database

NoSQL systems is over
🡪🡪 Specialized data systems

10
11
 The era of Big Data
Idea: make existing massive, unstructured data amounts usable

• Structured data (DBs)

• Log files
Sources
• Documents, Texts, Tables
• Images, Videos
• Sensor data
• Social Media, Data Services

• Statistics, Cubes, Reports

• Recommender
Analyst, Data Scientist, • Classificators, Clustering
Software Developer • Knowledge

12