2.7.4 · D1 · Maths › Statistics & Probability — Intermediate › Box-and-whisker plots — quartiles, IQR
Agar aap apna data chhote se bade order mein lagaao , toh us line mein positions — beech wali, aur quarter-marks — centre aur spread ke baare mein sab kuch bata deti hain, bina kisi ek pagal value se behe jaaye. Ek box plot sirf un position-marks ki picture hoti hai.
Is page mein assume kiya gaya hai ki aapko kuch bhi pata nahi . Parent note Box-and-whisker plots — quartiles, IQR padhne se pehle, aapko ideas ka ek chhota sa toolkit chahiye. Hum har ek ko zero se banate hain, usse ek picture se jodte hain, aur batate hain kyun is topic ko yeh chahiye . Baad ka har idea sirf pehle waale ideas pe tikta hai.
Symbols se pehle, raw material.
Definition Data set aur ek value
Ek data set sirf numbers ka ek collection hai jo aapne measure kiya — heights, incomes, test scores. Har ek akela number ek value hai. Hum aam taur par likhte hain ki kitni values hain ek letter mein, aksar n .
Picture: sochiye har value ek manka (bead) hai. Aapka poora data set muththi bhar loose mankyon ka hai.
Intuition Yeh word kyun chahiye
Baaki sab — median, quartile, spread — is mankyon ke dheer pe operations hain. Agar hum "dheer" aur "kitne manke n " naam nahi de sakte, toh hum jo kuch bhi us pe karte hain woh describe nahi kar sakte.
n hamara sabse pehla symbol hai.
n
n ka matlab hai data set mein kitni values hain unka count . Agar aapka data 2 , 3 , 4 , 5 , 7 , 9 , 12 hai, toh n = 7 kyunki saat numbers hain.
Is poore topic ka sabse important move.
Figure dekho. Upar, manke ek ulajhi hui dheer mein hain (unsorted). Neeche, wahi manke ek queue mein khade hain, sabse chhota left pe, sabse bada right pe. Us rearrangement ko ascending sorting kehte hain.
Definition Sorted (ascending) data
Ascending sort karna matlab hai values ko is tarah rearrange karna ki har ek pichle wale se ≥ ho — pehle sabse chhota, aakhir mein sabse bada. Kuch add ya remove nahi hota; sirf order badalta hai.
Intuition Kyun topic is ke bina shuru nahi ho sakta
Quartiles line mein position se define hote hain ("jo ek-quarter raaste pe hai"). Ek position tabhi kuch matlab rakhti hai jab sab order mein khade hon. Ulajhi hui dheer mein, "beech wala" bakwaas hai. Isliye har box-plot calculation ka Step 0 hai: sort karo.
Common mistake Sorting positions badalta hai, values nahi
Sorting koi bhi number nahi badalta. Value 30 abhi bhi 30 hai; woh bas wahan jaati hai jahan use line mein hona chahiye. Log kabhi kabhi darte hain ki sorting "information kho deta hai" — yeh kabhi nahi hota.
Yeh subtle hai aur iske liye apni picture worth hai.
Sorted queue mein har manke ke paas do alag numbers hote hain:
uski position — woh line mein kahan khada hai : 1st, 2nd, 3rd, … (neeche line ke yellow mein dikhaya gaya).
uski value — woh kaunsa number hai (manke ke andar blue mein dikhaya gaya).
Definition Position (rank)
Kisi value ki position (ya rank ) uska sorted queue mein jagah hai, left se 1 se count karte hue.
Intuition Unhein alag kyun rakhein
Quartiles pehle ek position choose karke (jaise "position 4") aur phir wahan ki value padhke milte hain. Dono ko confuse karna — ek position number ko data value ki tarah treat karna — classic beginner galti hai. Yellow ruler (positions) aur blue beads (values) ko alag mental drawers mein rakho.
Ek box plot queue ko 2 1 , 4 1 , 4 3 pe mark karta hai. Isliye hume bilkul clear hona chahiye ki queue ka ek fraction matlab kya hota hai.
Definition Raaste mein ek fraction aage
Queue mein "Aadha raasta " matlab woh point hai jahan dono taraf equally manke hain . "Ek-chauthai raasta " matlab har chaar mein se ek manke aapke left mein hai, teen right mein .
Picture: sorted queue mein kharo. Tab tak chalo jab tak theek aadhe log aapke peechhe hon — woh jagah halfway mark hai. Sirf ek-chauthai bheed tak chalo — woh quarter mark hai.
Intuition Average nahi, fractions kyun?
Hum middle ko saari values ka average lekar describe kar sakte hain. Lekin average har number add karta hai , isliye ek giant value usse kheench leti hai. Ek fraction-of-the-way position sirf order ki parwah karta hai, size ki nahi — isliye ek giant sirf "line mein aakhri insaan" hai, baaki kisi se zyada powerful nahi. Yahi poora reason hai ki position-based measures robust (dekho Outliers and Robust Statistics ) kyun kehlaate hain.
Ab hum halfway mark ka naam de sakte hain.
Definition Median, likha jaata hai
Q 2
Median sorted data ki halfway position pe value hai. Iska symbol Q 2 hai (doosra "quartile", agle mein explain kiya gaya). Lagbhag 50% data uske neeche hota hai.
Isko kaise dhundhein yeh depend karta hai ki n odd hai ya even — kyunki "bilkul beech" tabhi exist karta hai jab count odd ho.
n odd (upar wali row, 7 manke): ek akela beech wala manka hai. Uski position hai 2 n + 1 = 2 7 + 1 = 4 , isliye median 4th value hai.
n even (neeche wali row, 6 manke): koi single manka beech mein nahi hai — centre do mankyon ke beech padata hai (3rd aur 4th). Median unka average hai.
n even ho toh do middle values ka average kyun?
Even count ke saath, sahi "centre point" genuinely do mankyon ke beech gap mein hai. Kisi ek ko akele chunna thoda left ya right jhukaa dega. Unhe average karna bilkul gap mein land karta hai — honest beech. Aur dekho Median and Measures of Central Tendency .
"Quart" matlab chaar . Quartiles woh marks hain jo queue ko chaar equal-count groups mein kaatte hain.
Definition Lower aur upper quartiles
==Q 1 == (pehla / lower quartile): ek-chauthai raaste pe value — lagbhag 25% data uske neeche. Yeh lower half ka median hai.
==Q 3 == (teesra / upper quartile): teen-chauthai raaste pe value — lagbhag 75% neeche. Yeh upper half ka median hai.
Q 2 ke saath milke, teen cuts Q 1 , Q 2 , Q 3 chaar equal parts banate hain.
n /4 tak directly count karne ki jagah "har half ka median" kyun?
n /4 jaisi raw position aksar do mankyon ke beech padti hai aur alag textbooks us gap ko alag tarah patch karte hain. Clean, unambiguous rule hai: median pe split karo, phir har half ka median lo. Aap already jaante hain median kaise lena hai (§4), isliye yeh ek tool ko do baar reuse karta hai — kuch naya nahi seekhna.
Common mistake Kya median halves mein rakhte hain?
Agar n odd hai, toh median ek specific manka hai jo kisi bhi half mein belong nahi karta — usse drop karo, phir baaki split karo. Agar n even hai, toh koi single median manka nahi hai, isliye dono halves mein pehle se exactly n /2 manke hain. Yeh detail aapka Q 1 , Q 3 decide karta hai, isliye kabhi skip mat karo.
Yeh teeno symbols exactly 25th, 50th, aur 75th Percentiles and Quantiles hain — quartiles sirf nice quarter-marks pe percentiles hain.
Parent note ka spread measure hai Q 3 − Q 1 . Subtract kyun karte hain?
Definition Number line pe subtraction as gap
Number line pe, b − a a aur b ke beech distance (gap) measure karta hai. Bada difference → woh zyada door hain → data zyada spread out hai.
Intuition Yeh particular gap kyun?
Poora range (max − min, dekho Range and Spread ) do sabse extreme mankyon ke beech gap measure karta hai — sabse kam bharosemand wale. Q 3 − Q 1 iske bajaaye quarter-marks ke beech gap measure karta hai, dono taraf ke outer 25% ko ignore karke . Ek giant isse stretch nahi kar sakta. Isliye IQR robust spread hai.
Aakhri symbol unpack karna: number 1.5 aur word "fence".
Ek fence ek boundary line hai jo box se ek fixed distance pare rakhi jaati hai. Usse pare koi bhi manka unusually door wala flag hota hai — ek outlier .
Lower fence = Q 1 − 1.5 IQR , Upper fence = Q 3 + 1.5 IQR
Intuition IQR ko kisi number se multiply kyun karte hain (fixed number ki jagah)?
Humein "door" ke liye ek yardstick chahiye. IQR khud ek natural yardstick hai — yeh batata hai ki typical middle kitni wide hai. Har edge se "1.5 box-widths" chalna matlab distance aapke apne data ke spread ke saath scale karta hai: tight data ko tight fences milte hain, loose data ko loose fences. Choice 1.5 ek convention hai jo tune ki gayi hai taaki roughly 0.7% Normal Distribution data flag ho — rare lekin kabhi nahi nahi.
Fractions of the way along
Multiplier 1.5 and fences
Recall Kya aap bina dekhe har ek ka jawab de sakte hain?
Symbol n kya represent karta hai? ::: Data set mein kitni values hain unka count.
Quartiles dhundhne se pehle kisi bhi data ke saath aapko pehle (Step 0) kya karna chahiye? ::: Use ascending sort karo (chhote se bade tak).
Kisi value ki position aur uski value mein kya farq hai? ::: Position woh hai jahan woh sorted queue mein khada hai (1st, 2nd, …); value woh actual number hai jo woh hai.
Queue mein "ek-chauthai raasta" ka matlab kya hai? ::: Har chaar mein se ek manka aapke left mein hai, teen right mein.
Median Q 2 kya hai aur odd vs even n ke liye isko dhundhna kaise alag hota hai? ::: Halfway value; odd n ke liye yeh position ( n + 1 ) /2 pe single middle value hai, even n ke liye yeh do middle values ka average hai.
Q 1 aur Q 3 ek-ek phrase mein kya hain? ::: Q 1 = lower half ka median; Q 3 = upper half ka median.
b − a spread kyun measure karta hai? ::: Yeh number line pe distance/gap hai — bada gap matlab zyada spread out.
IQR formula likhो. ::: IQR = Q 3 − Q 1 .
Outlier fences set karne ke liye IQR ko (fixed number ki jagah) multiply kyun karte hain? ::: Taaki "door" wala yardstick data ke apne spread ke saath scale kare; har edge se 1.5 IQR door.